The case of romantic relationships: analysis of the use of metaphorical frames with ‘traditional family’ and related terms in political Telegram posts in three countries and three languages

1 Introduction

Romantic relationships define one of people’s most intimate aspects of life. They are associated with people’s gender identity, sexual orientation, spirituality, but also formal and institutional aspects that became social norms: marriage, monogamy, heterosexuality, and offspring. They are also associated with cultural aspects such as an “appropriate” age of marriage, expectation to be engaged in marriage and sometimes compulsory marriage, expectation for a young couple to give birth to children soon after marriage and so on.

Despite severe critique from feminist, anti-racist and anti-colonialist activists and authors, such as (Hooks 2021; Nair 2010; Roig 2023; Willey 2016), some forms of romantic relationships are still seen as more preferred over the others. Some behaviours within those socially preferred relationships are seen as more favourable than others. Powerful groups in the society take the right to decide and to judge about what is a “normal” relationship, and what is not. In the dataset used for this study and explained below, the “normal” relationship is often called “traditional family”. The notion of a “traditional family” in the dataset is used in different contexts and with different purposes.

Political news talk about romantic relationships in different ways, and topics on the agenda include gay marriages, abortions, child raising, sexual education at school, gender (in)equality and the role of social gender categories in all of these. However, media publications related to romantic relationships do not always work on family issues as such, but are misused for other purposes, mainly related to power. The objective of this is article is not to criticize the “traditional family”, although many things can be criticised, and multiple works cited here provide critical arguments. The objective of this article is to disclose and to question the instrumentalization of the concept “traditional family” for political and propaganda purposes unrelated to matters of personal relationships.

Figure 1 shows a news post from a mainstream state news channel in Belarus called ONT. The picture shows a former presidential candidate Sviatlana Tikhanovskaya, and the comment says literally “Tikhanovskaya in Warsaw: the meatball is having a nice life paid by Lithuanian taxpayers, she is traveling around in Gayrope”. The message contains several problematic labels. First, the meatball is referring to the presidential candidate herself, pointing to her profession before she became a politician, and inferring a mismatch between the person’s qualification and her role in politics. Second, the message points to the socially unacceptable way of living at the cost of taxpayers of the country that gave her shelter. And third, the word Gayrope, a compound of gay and Europe is used here and in Russian-language media in general to make Europe look negatively.

Figure 1:

ONT NEWS post on Telegram showing Sviatlana Tikhanouskaya walking on a street with negative textual framing.

Why do these things work? Especially, why is the word gay attached to the word Europe supposed to picture Europe negatively, make it look ridiculous or weird? Why does gayness of some Europeans appear in messages related to politics in Belarus? In fact, sexual orientation and intimate relationships of European people can be assumed to be their personal matters.

In theory they are personal, but in practice, emotions, language and intimacy are political, and they are used to shape culture. The work of modern feminist authors such as (Hooks 2021; Kurt 2021; Roig 2023; Solnit 2017; Stokowski 2018) demonstrates how emotions and intimate relationships have been politicized and used to manipulate and control populations.

Relationships between a man and a woman within a “traditional family” are not just romantic or economic; they are also about distribution and regulation of power, and thus, they are political (Kurt 2021; Roig 2023; Willey 2016). Feminist political and cultural studies argue that norms of monogamy and a “traditional” family are an instrument of resource distribution and a mechanism for a state-driven regulation of which forms of human social networks can exist. Married couples let the state acknowledge that their relationships count (Nair 2010).

For women, it was a long way to arrive at the “modern family” stage, where a woman can decide herself and study, work, vote, have a passport, drive a car, travel and be independent, have or not have children, have her personal interests. Also birth control still is a topic of hot debates and populist political slogans. Gender equality and sexual diversity discourses argue that gay marriage must be acknowledged by state to guarantee human rights for all types of romantic relationships. As a consequence, “the fight for recognition of same-sex marriage also testifies to the pervasive couple-centricity of our culture. The dyad, for so long opposite sex and now increasingly also same sex, is portrayed as the fundamental unit of love and family. It is a key structure used to try and gain what should be fundamental human and civil rights for all our citizens” (Kim 2014). Nevertheless, as long as the institution of marriage exists, it should be for everyone, as (Roig 2023) argues.

Although some authors claim that the traditional family was important for capitalism as a unit of power and capital distribution (Kurt 2021; Nair 2010), communist ideology also made use of the concept “traditional family” to control the society. While originally proposed by K. Marx and F. Engels, new forms of romantic relationships and child-raising communities would replace the “bourgoise family” (Gorecki 1972) recent studies of post-communism societies show that the opposite was the case in practice: preference for “traditional families” and marriages increased, especially with regard to the role of family in childcare and elderly care (Costa-Font and Nicińska 2023) and the family was declared to be “the basic unit of the society”. In any case, romantic relationships are subject to regulation under any of these political systems, and each of these political systems preferred married cis couples over any other form of relationships (Pongracz and Spèder 2008; Stankuniene and Maslauskaite 2008). The only difference among those political systems was the degree of how strictly all other forms of romantic relationships were discriminated against or sanctioned, and for which purpose.

Language is employed to craft emotional appeals and disseminate propaganda, shaping cultures and directing masses according to the desires of those in power. As Stokowski (2018) argues, politics has always been about things that have a lot to do with emotions. “Fear, disappointment, hope, anger, solidarity, longing, revenge, and security needs are quite subjective but effective matters. Rationality and knowledge of facts were never the main idea of politics.”^[1](Stokowski 2018).

This article analyses how intimate relationships are handled in political discourse in three countries and three languages and for which purpose. A new dataset from Telegram posts has been compiled for the purpose of this analysis and is described in detail in Section 4. First, this article discusses literature highlighting the relationships between emotions, culture language (especially linguistic metaphors and agency formulations) and propaganda (Sections 2 and 3). Further, it demonstrates on examples from the corpus how people’s most intimate areas such as romantic relationships are used to spread political agenda on modern social media (Section 4). Finally, this work outlines possible solutions and makes some recommendation for future regulations and research (Section 5).

2 Emotions, culture, propaganda and intimacy

Culture refers to the shared beliefs, values, practices, behaviors, and customs that characterize a group or society. It includes the language, art, music, cuisine, traditions, and other aspects of a community’s way of life. Culture is passed down from generation to generation and can shape individual and collective identities, social norms, and worldviews. Culture can vary greatly between different regions, ethnicities, religions, and other social groups (definition generated by ChatGPT turbo 3.5).

Propaganda refers to the spread of information, ideas, or opinions designed to influence people’s attitudes, beliefs, or behaviours in a particular way. Propaganda can be used to promote a particular agenda, ideology, or political message, often through the use of biased or misleading information, emotional appeals, and other persuasive tactics. It is often associated with political or military campaigns, but can also be used in advertising, public relations, and other forms of communication.

If propaganda shapes our beliefs and attitudes, how much of the culture is then passed from generation to generation? And how does the culture appear? If culture is about values and norms, and propaganda can change them, how do people know what values they have and what norms to use for their judgements?

Propaganda often makes use of emotional appeals and simplistic messages that can bypass critical thinking, regardless of an individual’s intelligence level. Furthermore, intelligent people may be more confident in their ability to discern fact from fiction, making them less likely to question the information they receive, which can make them more receptive to propaganda (Allcott and Gentzkow 2017).

As (Stokowski 2018) argues, post-factual as a term may have appeared in the last decade in relation with media campaigns of politicians, however post-factual as a way of mass manipulation was already described 500 years ago by Niccolò Machiavelli (1513) in his work The Prince: “and men are so simple, and so subject to present necessities, that he who seeks to deceive will always find someone who will allow himself to be deceived” (Machiavelli 1993: 84). One could argue that 500 years ago most people were illiterate, and mass media did not exist at all. However, as (Stokowski 2018) adds, not much changed after people received better access to literacy and education. For example, Max Weber criticised in his lecture Politics as a Vocation in 1919 that purely emotional means are used in politics. He suggested to call the existing state of affairs “a dictatorship resting on the exploitation of mass emotionality” (Weber 2004).

Although the majority of people 100 years ago did not have the same access to information as nowadays, the vast amount of information and opinions today makes it difficult for the majority to differentiate between facts and lies, and our inherently emotional human nature makes it difficult to distinguish between emotional appeals and rational argumentation in political discourses. Although the modern information media-driven society is sometimes called “post factual”, using emotional appeals and so-called “alternative facts” is nothing new. It was not much different already almost a century ago, as (Stokowski 2018) notes. A German philosopher Theodor W. Adorno (1973) observed common characteristics of modern fascist movements, and argues that they all twist facts deliberately, and succeed with those who are not well informed. Complex societal problems lead to a common state of insecurity, and this combined with disinterest in facts provides a nurturing ground for reactionary mass movements, which are always ethnocentric and anti-intellectual (Adorno 1973).

The difference between the time of Adorno and Weber and now is in the use of information technology and algorithms to distribute information to the right audience with speed and at scale that was not imaginable by that time. To compare (Jones 2020), illustrates how misinformation about homosexual individuals was spread in 1987 and how it is spreading now with social media. In 1987, Soviet actors spread narratives that the US government injected young gay men in other countries who were the source of the AIDS crisis. These narratives were distributed in 80 countries through 200+ printed periodicals in 25 languages. In 2015, in just two days (Jones 2020), received 778,464 views and 65,510 clicks on their 3,519 memes on Facebook. Technology enormously facilitates and speeds up precisely tailored distribution of messages to the target population. Social media provide tools to deliver messages to people who would most likely engage with them (like, repost, comment). This allows for controlled and intentional mass opinion manipulation and belief injection. With time, new beliefs become a new culture. A former Cambridge Analytica employee and whistle-blower explains how it works:

When a communicable disease threatens a population, you immunise certain vectors first … as they are most susceptible to infection. Then nurses and doctors, teachers and bus drivers, as they are most likely to spread a contagion through wide social interaction, even if they do not succumb to the disease themselves. The same type of strategy could help you change culture. …you would first identify which people are susceptible to weaponised messaging, determine the traits that make them vulnerable to the contagion narrative, and then target them with an inoculating counter-narrative in an effort to change their behavior. (Wylie 2019: 56)

The search for technological and technology-assisted solutions to counter the spread of propaganda led to emergence of new academic venues such as FAccT conference^[2] and NGOs such as Institute for Strategic Dialogue.^[3] In addition, automated identification of metaphors in political discourse has been studied with computational and machine-learning means (Picard and Stammbach 2022; Rocco 2023). However, computational approaches have their limits, as this article discusses in Section 5.

This article draws relationships between language, politics, emotions, culture and new technologies. For example, in Belarussian presidential election campaigns 2020, the supporters of Alexander Lukashenko use tradition, stability and peace in their slogans as the opposite to change which would, according to them, lead to chaos, insecurity and conflicts. As (Roig 2023) argues, “tradition” sounds good, it gives a feeling of belonging and community. A “traditional” family, however, is actually a patriarchal family in which women have less value than men, and gender binarism is the necessary prerequisite for the implementation of the gender-based inequalities. The culture of “traditional” families has been created in the interest of capitalism and continued to be exploited by other political systems. It is also a European-centric concept that was exported to all other parts of the world while other, historically existing forms of relationships and other genders incompatible with the understanding of the family as a union of a man and a woman were declared uncivilized, primitive and undesired (Roig 2023). The effect of the word “traditional” is that it makes the patriarchal family look good even for those who are exploited. This article seeks explanation for the complex relationships among culture, emotions, politics and language with the concept of a “traditional family” as an example.

3 Language, perception, emotions and politics

Cognitive linguists, other researchers, and authors agree that the language we use influences our perception (Boroditsky 2009). The point of disagreement is, to what extent language influences our perception (Gümüsay 2020). Research has shown that language features such as grammatical gender of person nouns and metaphors and attributes used to describe nouns influence how people threat those subjects (Boroditsky 2000). For example, in languages that have grammatical gender markers in nouns describing persons’ occupations, such as Russian and German, text authors can choose whether they use some inclusive forms of occupation names addressing persons of all genders, or just a male form (simplicity and readability argument). Research shows that readers’ perceptions of which genders can represent various professions highly depend on whether inclusive or solely male forms of profession titles are used (Frauenknecht et al. 2021). Grammatical gender of nouns that are not related to persons influences the ways people describe those things (more stereotypically masculine adjectives for male things such as key in German and more stereotypically feminine descriptions for female things such as bridge in German). Further, an increased use of inclusive language is connected to more tolerance towards immigrants and less homophobia (Frauenknecht, et al. 2021).

Also, metaphors are shown to be a strong linguistic instrument capable of influencing people’s perception of the world (Chkhaidze et al. 2021). For example, according to (Hendricks et al. 2018), metaphors can modify the way people think even about such prominent, highly politicized, and media-saturated issues as immigration. In their studies, political messages differed in just one word describing immigration as ‘boost’, ‘increase’, ‘invasion’ or ‘flood’ (Thibodeau et al. 2017). The participants’ predictions about the effects differed significantly depending on the metaphor used. Studies about metaphors and perception show that metaphors can act covertly in organising people’s beliefs. As (Thibodeau and Boroditsky 2011) point out, people do not recognize and not perceive the influence of metaphors on their actions, however, even single-word metaphors can have a powerful influence on how people suggest solving social problems and how they gather information to make decisions. While people’s arguments involve statistics and numbers, the experiments by (Thibodeau and Boroditsky 2011) show that metaphors have a much stronger influence on the decision-making process than so-called objective reasons.

A series of experiments in 2011 and 2014 conducted by (Thibodeau and Boroditsky 2015) shows that a metaphorical frame is especially powerful if introduced early to the issue in focus. In this way, the metaphor used can actively shape the mental representation of the issue and, consequently, the proposed solutions. In their experiments, framing crime as a beast received support for enforcement-based solutions in participants while framing crime as a virus led to support for reform-based solutions (Thibodeau and Boroditsky 2015). also found that norms influence how we categorize these solutions. For instance, the solution to “increase neighbourhood watches” was classified as enforcement-related in 2011 but it was classified as reforms-related in 2014. These findings demonstrate the importance of data sharing and replication studies and also that the shift in human perception on concepts due to metaphors is not static, and language change goes along with societal changes and norm changes.

Sometimes, the right language to describe things must be found first. We cannot speak about issues if we do not have the right words for them. As (Solnit 2017) argues, with the feminist movement, language to describe problems caused by patriarchal power abuse had to be invented: terms such as ‘rape culture’, ‘domestic violence’ and ‘sexual abuse’ have not existed before, but they emerged from the protest movement against those things, and we use these terms now naturally. Nevertheless, it does play a role how we speak about the issues, not just the fact that we can name them. For instance (Roig 2021), suggests using the term ‘violence in public spaces’ instead of the popular term ‘street violence’ because the word ‘street’ in collocations like ‘street dance’ and ‘street style’ is associated with people of colour as agents, while ‘violence in public spaces’ includes all people as potential agents and victims. Further, the study by (Schnepf and Christmann 2023) shows that metaphors used to describe cases of femicide influence cognitive responses to crime. If a crime case was framed as a ‘love killing’, the killer earned more empathy, especially from male subjects, than when framed as a ‘murder’. Until recently, marital rape did not count as crime; it was handled as private family space and personal matters, and it remains like this in some communities (Adinkrah 2011; Mahoney and Williams 1998). This perception may be a consequence of the “traditional family” – a construct in which a man owns a woman’s sexuality, is entitled to sexual relationships with his wife, and a wife should not only engage consensually but also enjoy sexual relationships with her husband, making rape unnecessary (Mahoney and Williams 1998; Roig 2023). Although almost 40 % of women in various shelters in US report experiencing forced intercourse from their partners, they prefer not to use the word ‘rape’ for it (Finkelhor and Yllo 1983).

Metaphors facilitate emotional framing and successful fallacious arguments in political reports, as our earlier work shows (Höhn et al. 2023). However, most intuitive linguistic choices of journalists to describe particular events can be explicitly censored and prohibited. For example, a former employee of Russian state TV channel Rossiya 1 explains how she had to replace all mentions of “protesters” in her reports by “spiritual heirs of Stephan Bandera”. ‘I was writing the text at that time and referred to the protesters as protesters, and my immediate supervisor, in a stern tone over the phone, told me not to write like that anymore and to refer to them as ‘spiritual heirs of Stepan Bandera.’^[4] Our earlier work shows how metaphors were used by Russian media to depict Ukrainians as nazis (Höhn et al. 2023) and how such beliefs are cultivated and reinforced by using more and more extreme metaphors over time (Höhn et al. 2021, 2022]).

Family-partnership-love metaphors are used in public discourse to describe relationships among political entities such as states and nations. As (Musolff 2006) shows with the analysis of a bilingual German-British corpus of metaphorical descriptions of political entities, both cultures, British and German share not only the use of family metaphors for political entities, but also cultural stereotypes, such as solidarity, male-centredness and marriage as privilege that implies rights and duties. Thus, family metaphor activates expectations of a family-like behaviour of non-family-formed entities such as states and nations. Thus, whenever family metaphors are used to describe political entities, “traditional family” is supposed to be understood.

Agentive formulations are also used to make this framing more effective. For example (Stanojević et al. 2023), among others discuss the effects of climate change framing as ‘we are moving towards a catastrophe’ versus ‘the catastrophe is approaching’. Although the effect of the catastrophe can be expected to be the same, the former formulation makes readers feel more in control and therefore, less in danger. Agency formulations seem to be a factor influencing our perception. Even if people have enough evidence to judge about events, language used to describe those events affects most people’s judgments about whom to blame and severity of suggested consequences (Fausey et al. 2010). Especially, agentive formulations lead to more blame and requests for higher penalties than non-agentive formulations, even if study participants have access to video footage of the situation about which they need to judge. Experiments in (Fausey et al. 2010) suggest that direct manipulations of the agency formulations lead to changes in eye-witness memories: agentive formulations help people to remember the events better than non-agentive. Further, preference for agentive formulations is influenced by culture, as experiments with English and Japanese speakers show. As demonstrated in our earlier work, agency formulations are one of the instruments of bias in political news (Höhn et al. 2023). Altogether, agentive formulations lead to better memorization of events, and higher blame and requests for stronger punishment, even if the truth value of those formulations is questionable.

Metaphors help people to think more efficiently. Some metaphors are more popular than others. For instance, warfare metaphors were used in 17 % of all articles published in Time Magazine at least once and in 15 % of all articles in Newsweek between 1981 and 2000 (Karlberg and Buell 2005). The study of war metaphors in public discourse by (Flusberg et al. 2018) suggests that warfare metaphors are so frequent because they activate very familiar frames. As the authors argue, metaphors are effective when they activate a knowledge structure, the knowledge is well-known to the speaker, and the comparison between the source and the target is appropriate in the community culture. The war metaphor meets all these criteria. The authors note that the interpretation of the particular instance of a war metaphor is shaped by multiple factors, such as culture and context of use. For instance, the war metaphors analysed in (Höhn et al. 2021) refer to the cultural and historical aspects of WWII on the Belarussian territory (usually called ‘Great Patriotic War’ in Belarus in Russia). The limitations and dangers of the war metaphor are in its dividing power of the world into ‘us’ and ‘them’, those who fight against an ‘enemy’ and the ‘enemy’ (Flusberg et al. 2018).

Metaphorical framing influences the ways we cope emotionally with negative events in our lives. The study by (Hendricks et al. 2018) shows that framing cancer as a battle is associated with guilt while framing cancer as a journey is associated with making peace with the condition. Also conflicts in romantic relationships are perceived and handled differently when framed differently. The study by (Lee and Schwarz 2014) shows that relational conflicts hurt more if a romantic relationship is framed as a perfect unity than if the frame of a journey with ups and downs is used. Framing a relationship as a journey elicits expressions such as “taking a journey together”, “look how far we’ve come”, “we’re spinning our wheels”, “the marriage is going nowhere”, “we’re at a new place in our relationship”, “we have to go our separate ways”, and “it’s been a long and bumpy road” and framing an argument as a war is correlated with expressions like “he attacked my point,” “I won the argument,” “his position was indefensible,” and “he shot down my argument” (Robins and Mayer 2000). Robins and Mayer (2000) also report that framing a relationship as a war leads to increased participants’ preference for guarded communication than framing a relationship as a two-way street. As (Musolff 2017) argues in his analysis of the parasite metaphor in debates about immigration in UK, the use of strong metaphors in political communication should be considered strongly intentional. Those who compare any social groups with bloodsuckers know what they are doing (Musolff 2017), notes also referring to Nazi discourse about oppressed social groups (Musolff 2010).

Modern feminist authors such as (Hooks 2021; Roig 2023) among others criticize that what they call ‘compulsory monogamy’ and ‘compulsory heterosexuality’ is framed as a search for your second half (implying that you cannot be a complete person without a partner of the opposite gender) (Wawrzyniuk 2023). argues that, given the language’s formative power on our perception of reality, we can use the power of metaphors to maintain or to reject gender ideologies. For example (Wawrzyniuk 2023), shows how women-objectification metaphors are used in stand-up comedy to break socially constructed gender stereotypes and norms. Thus, metaphors not only influence our thinking and our actions, but also our emotions and feelings, even regarding such personal issues as romantic relationships. Consequently, they impact our habits, our beliefs and finally, our culture.

Persuasive and political communication is full of metaphors. Extended metaphors were found even more effective in influencing policy attitudes in people than simple metaphors (Thibodeau 2016). However, research shows that extreme metaphors can have an opposite effect: in case of immigration, extreme metaphors alerted the study participants and led to more empathetic measure suggestions (Hart 2021). Different methods used in metaphor research can also lead to different results. As (Boeynaems et al. 2017) report in their critical literature review, the research on the impact of metaphor in politics is still fragmented.

In social media, posts often contain visual materials such as pictures, GIFs and videos. What we perceive by looking at a picture is determined not only by physical properties of the picture, but also by our perception history (Lupyan et al. 2020). As Lupyan et al. (2020) discuss, language influences human visual perception in classification and discrimination tasks. Language also “seems to modulate perception in the moment” (Lupyan et al. 2020) and can provide a categorical expectation with which visual input is processed. The reader is invited to look one more time at Figure 1, a totally innocent picture of a politician on a street, and to pay attention to the ways linguistic framing is influencing their interpretation of the picture. The linguistic framing in the message regulates the categorical expectation and manipulates the viewer’s interpretation of what is actually in the picture.

To make a metaphorical frame work, it is sufficient to mention just some actions or properties of a category to restore the entire category. This phenomenon can be explained using membership categorisation analysis (MCA) (Schegloff 2007; Sacks and Jefferson 1995) which has been shown effective in critical discourse analysis (Freiberg and Freebody 2009). As (Freiberg and Freebody 2009) demonstrate, each social category is stored in our social category cognition as an aggregate of features that are tied to category labels at different degrees and include what Sacks called category-bound activities (Sacks 1991):

1. Constitutive features are type-embedded and criterial for that category, and must be observable;

2. Tied features are criterial for that category under certain conditions, and will be definitely observable when those conditions occur; can generate category membership under those conditions;

3. Occasioned features are not criterial but might be made criterial under certain conditions; not sufficient to infer about category membership.

These features are important because the observers (or readers of a news text) categorise people not only via explicit labelling but also via descriptions of their actions or attributes. Thus, in the analysis of the metaphorical framing in the following sections, not only direct labels will be taken into account, but also constitutive and tight features of categories, that the readers are supposed to recognise from those descriptions.

4 Cultivation of romantic relationships in language and political agenda

Not only are institutionalized forms of romantic relationships the subject of political debates, moral norms, and legal regulations, but also topics related to relationships such as a person’s sexual preferences and gender identity.

Figure 2 shows a post by BELTA on 5.12.2022 saying “Russia’s parliament issued a law prohibiting propaganda of LGBT, paedophilia and transgender.” After this, any information related to non-cis gender identity or non-heterosexual practices were declared “propaganda” and could be legally sanctioned, and gender queerness has been put in the same category of crime as paedophilia. While this picture shows only one neutral-looking post from a state media channel, many discussions occur in the comments on those posts and generally within unregulated media – instant messengers, including Telegram, Gab, Parler, Mastodon and 4chan, but also on Instagram and TikTok. So-called “dark platforms” offer free space for hate and toxicity, but also for covert propaganda (Höhn, Mauw & Asher 2022).

Figure 2:

Telegram post announcing the new law in Russia that prohibits ‘LGBT-propaganda, paedophilia and transgender.’

Telegram was launched in 2013 as a space for safe communication. From a mainly Russian-speaking anti-government messenger in the beginning, Telegram became influential platform used by Russian regime to spread propaganda (Salikov 2019; Wijermars 2022). In Europe, and specifically Germany, Telegram became the platform for spreading false information about pandemics, conspiracy theories and anti-European thinking (Kiess 2023; Zehring and Domahidi 2023). In the US, Telegram became popular in extremist circles (both far-right and far-left) (Blas et al. 2024; Davey and Weinberg 2021; Walther and McCoy 2021) as they were banned from established social media platforms (Rogers 2020).

For this study, a new corpus was compiled from news posted on nine Telegram channels in three languages: Russian, English and German. Russian-language channels include two Belarus state media ONT NEWS and BELTA (state-owned mainstream media), and one Ukrainian outlet Ukraina Online (one of the three most popular Ukrainian Telegram channels). German-language channels include AfD Brennpunkt, AfD im EU Parlament and AfD Kompakt (right-wing anti-European conservatives). US English channels include Donald J. Trump, Steve Bannon’s War Room Pandemic (banned from most news media; conservative politician, board member of Cambridge Analytica and Breitbart News), and Nicholas J. Fuentes (the leader of the Groypers, a far-right white nationalist group). All these channels were selected because they emphasize the importance and the value of “traditional families” for the society in the countries to which they belong, however, they frame the “tradition” differently. For this study, posts were selected from the above-mentioned channels that were broadcasted between 1. August 2020 and 31 December 2022 and contain keywords traditional family, family values, LGBT, abortion, gay marriage and gender. The final dataset contained around 500 posts from all channels in total. The dataset will be shared upon request. All Russian and German terms and examples are translated into English by the author of this article. The dataset was analysed qualitatively focusing on metaphors used along with the above keywords. The analysis also covered the warfare framing in the data.

The study is methodologically informed by the conceptual metaphor theory (Lakoff and Johnson 2020) and membership categorisation analysis (Freiberg and Freebody 2009). For the analysis, all occurrences of the terms of interest traditional family, family values, LGBT, abortion, gay marriage and gender were collected in three languages and all channels together with their co-occurring terms but within borders of syntactic units (noun and verb phrases) that are referred to as collocations. For all collocations it has been analysed, whether they are metaphors themselves or make use of constitutive features of conceptual metaphors and make those concepts recognisable in this way. The following subsections section present the results of the analysis.

4.2 Gender and LGBT concepts

Especially in the German part of the dataset, the concepts of “gender” and “LGBT” are frequently used interchangeably, however, in the first part of the German data timeline, “gender” is mostly used in relation to grammatical forms in the German language and against gender-inclusive use of language. The transition from the concept of ‘gender’ related to gender inclusivity in language to ‘gender’ as self-identification occurs over time, after which the term ‘LGBT’ is predominantly used. However, the metaphors are still the same: “insanity”, “ideology”, “propaganda”, “trap”.

Interestingly, channels in all analysed languages mostly make use of the same metaphors in relation with gender and LGBT concepts: propaganda, dictatorship, agenda and their synonyms while the overlap in metaphors used in Russian and English-language channels is even bigger: virus, anti-Christian, paedophile. Typical in all Russian-language media is the use of the LGBT label to discredit individuals by associating homosexual orientation with corruption and behavior incompatible with official duties, see for instance https://t.me/UaOnlii/13245.

We can observe differences in agency attribution related to LGBT terms: while Belarussian channels accuse EU of “LGBT ideology”, German and English-language posts accuse left-wing and green parties of “LGBT indoctrination” to index “us” and “them”. However, both Belarussian and US channels use the same false claims in their texts, such as “Gays are 3 % of the population but 40 % of the pedophiles” (Fact check: this claim is false!^[5]). While Donald J. Trump channel draws connections between homosexuality and AIDS (see the reference to an old Soviet propaganda campaign (Jones 2020)), other US channels in the dataset are more aggressive in their lexical choices.

The metaphors used by the Ukrainian channel disclose homophobia deeply rooted in the society and orientation to European values at the same time. This suggests that the analysis in the change of metaphors can be an indicator of cultural changes. In fact, Figure 3 shows how the narratives about the same topic can change in the political discourse over quite a short period of time. Figure 3 A) announces a new law proposal that would sanction “LGBT propaganda” in July 2020, and in December 2021, the official administration building was illuminated in LGBT colours in support of the International Day of Human Rights. Changes in political orientation of the government and the country led to changes in language, and later, it can lead to changes in culture (Table 1).

Figure 3:

Posts on 27.07.2020 (left) and 11.12.2021 (right) illustrating changes in framing of LGBT concepts in messages on the same channel. A) Propaganda of homosexuality in Ukraine will be liable to prosecution. B) The building of the Kiev city administration was illuminated in LGBT colors because of the International Day of Human Rights.

Table 1:

List of collocations used to frame ‘gender’ and ‘LGBT’ concepts.

Russian (Belarus)	Russian (Ukraine)	German	English
Пропаганда гомосексуализма	Пропагандa гомосексуализма	Gender-Gaga	LGBT agenda
ЛГБТ-пропаганда	видимости и принятия ЛГБТ	Gender-Irrsinn	Gay agenda
Гомо-гнездо	лгбт-любовь	Gender-Quatsch	Anti-gay ad
ЛГБТ-диктатура	дискриминация ЛГБТ	Gender-Falle	Fixing gayness
“прогрессивная” толпа цветных феминистов и ЛГБТ-активистов	продвижение однополой любви среди детей	Gender-Mainstreaming-Politik	LGBT / Satanic marchers
ЛГБТ-вирус	ЛГБТшник	Gender-Fanatismus	LGBT propaganda
Гейропа	«ЛГБТ-квотa»	Blödsinn	“LGBT positive”
ЛГБТ-идеология	защитa прав ЛГБТ	Ideologischer Fake	pedophile LGBT
Извращение	права ЛГБТ	Gender-Ideologie	LGBT indoctrination
участие в ЛГБТ-оргиях	развращение умов малолетних	Lebensformen gesellschaftlicher Minderheiten	Queer theory and pedophilia
евросодом	показательный каминг-аут	Gender-Wahnsinn	normalising pedophilia
	квир-банкир	Gender-Agenda	moral corruption
начало конца всего человеческого		LGBTIQ- und Gender-Ideologie	promote queer-theory
пропаганды ЛГБТ, педофилии и смены пола		LGBT-Kult	promote lesbian parents
		LGBT-Propaganda	ban gay marriage
		Werteimperialismus und Diversity-Ideologie	ban gender studies
		LGBT-Ideologie	gender ideology
		Verstörende LGBT-Pläne	homosexual ideology
		LGBTIQ-Kult	bestiality pride march

5 Discussion and conclusions

Argumentation by “tradition” has many faces, such as appeals for return to “traditional” role of women in families (including domestic violence, misogyny and dehumanization of women), “traditional” sexual orientation (including homophobia, transphobia and compulsory heterosexuality) and “traditional” marriages (compulsory monogamy with a woman subordinate to a man and compulsory offspring also subordinate to their father). Appeals to “traditions” usually build upon a human need for security, stability, and are usually the opposite of change. “Traditional family” arguments were also used to support the “security & stability” messages, and all forms of gender identity and sexual orientation that are different from monogamic unions of a cis-man and a cis-woman were used as examples of chaos and the “West’s” attempts to destroy Belarus in some way (destroy culture, tradition etc.). Interestingly, the “Western” channels analysed in this article also accuse their political opponents of attempts to destroy traditions. Thus, the main function of the “save traditional family” arguments is not tradition preservation but drawing the line between in- and out-groups, constructing “us” and “them” in order to control the society.

Political agenda and political narrative are shaping our culture. Protest movements such as feminism also shape our culture. Today, when we speak about “traditional family” rarely someone would imagine a tradition in which women have no voice and no rights, as it was just 100 years ago. And if marriages are unions of two equal persons, marriages of two persons of the same gender may sound reasonable. However, as (Solnit 2017) puts it:

The idea that a marriage could unite two people of the same gender only arose because feminists broke the institution of marriage from the hierarchical system of which it had been a part, reinventing it as a relationship between equals. Much suggests that those who perceive the opening of marriage as a threat feel threatened by the idea of an equal relationship itself.^[6]

While public discussions of institutionalized forms of relationships are a normal part of democratic societies, authoritarian societies and conservative right-wing parties in the dataset used for this work instrumentalise these discussions to promote their own agenda and spread propaganda. While other political actors may use metaphors in the same ways but for own purposes, the aim of this article was to uncover the mechanisms how such metaphorical framings work.

Digital technologies, as discussed in Section 2, facilitate the spread of propaganda on the one hand and offer possible solutions on the other hand. Two technical solutions are frequently used: automated detection of misinformation and toxic language, and content moderation by human operators. However, both solutions have their limitations.

Automated detection is usually based on lexical patterns in language, which does not reflect the variety of linguistic means used for propaganda purposes and also cannot deal with language variation and change (Höhn et al. 2023). The study by (Dias Oliva et al. 2021) uses Perspective, an AI technology developed by Jigsaw (formerly Google Ideas), to measure the levels of toxicity of tweets from prominent drag queens in the United States. The experiments show that Perspectives considered a significant number of drag queen Twitter accounts to have higher levels of toxicity than white nationalists (Dias Oliva et al. 2021). Cultural markers are identified as “toxic” language also in the study by (Zhou et al. 2021).

Content moderation relies on human moderators who read and filter harmful content. Companies are criticized for their colonial approaches to content moderation (Sarkar 2023). For example, OpenAI has been criticized for underpaying Kenian workers for reading and filtering hate speech from ChatGPT training data (Muldoon and Wu 2023). Further, untransparent practices of content moderation can be discriminatory for marginalized users of social networks (Thach et al. 2022). As an effect of “debiasing” practices, marginalised communities and their linguistic markers disappear from large language models and generated text (Xu et al. 2021). The understanding of marginalised communities in Large Language Models is very limited, simplistic, overgeneralised and stereotypical (Gadiraju et al. 2023). Large language models amplify standard language ideologies and promote only particular worldviews. This is a warning, because these worldviews can vary, and fine-tuning models to promote views that endanger specific communities while overprivileging others is technically quite simple. Such world views include but are not limited to monogamy, cis-sexuality and male-female gender dichotomy ideologies.

This observation leads us back to Adorno’s words that “all fascist movements target the uninformed” (Adorno 1973). In the context of the current information technology landscape, it is hard to call people “uninformed”. People are informed, but their choices of information sources are often not their own, and their beliefs are frequently not their own either.

While technological solutions can provide some help, other strategies might be more effective. Because the problem is very complex, potential solutions need to include instruments addressing information literacy, policy-making with clear definitions of liability and accountability. On the literacy level, we need to create awareness of emotional persuasion and other fallacies early in schools. This is hardly possible in countries such as Belarus where schools are one of the most important propaganda distributors, but in democracies such as Germany and USA it must be doable. Calling misogyny, homophobia and transphobia a “tradition” and “culture” has nothing to do with freedom. On the contrary, such appeals to “traditions” attempt to cut on human rights and amplify power inequalities between white cis-man and all the others. On the policy and liability level, authorities need to enforce certification of methods, algorithms and models and work with affected communities and end-users of the models to create effective and accepted policies. They also need to define clear liability rules for propaganda crimes and discrimination.

In addition, critical engagement with habitual and historical concepts such as ‘tradition’ and ‘family’ can be eye-opening. A “traditional family” is in fact a patriarchal construct enforcing cis-male dominance. All supporters of “traditional families” can educate themselves, especially its non-male supporters. Because the only category that benefits from a “traditional family” is a man. It needs to be emphasized that it is not an individual man, but the entire social category, with all its political, structural and systemic power and constructed superiority over all other categories.

Finally, to make technological support for solving these issues, we may need to rethink the design paradigms in the language technology production pipelines. We may need to renew the theoretical grounding of language technologies, replace outdated language theories by newer ones, and find clear language to describe things that are going on. For instance, we can name these calls for traditional family values “promotion of patriarchy” or “anti-feminism war”. This step is probably the one to start with because the language we use makes a difference for how we perceive all other steps, and what we perceive as a solution.