Self-censorships in Ukraine: Distinguishing between the silences of television journalism

Introduction

This article takes an empirically grounded approach to self-censorship to demonstrate that this phenomenon not only takes many different forms that function according to distinct logics, but also that these forms can coexist within one empirical moment. The same professional group, indeed the same person, can engage in different kinds of self-censorial practice in one moment in time. Building on insights developed empirically in my study among TV journalists in Kyiv, Ukraine, I argue that scholars should pay more attention to the heterogeneity of empirical forms of self-censorship, and to the ways in which these occur in the same context. Contributing to existing scholarly understandings of self-censorship (Bourdieu, 1991; Jungblut and Hoxha, 2017; Lee and Chan, 2009; Mortensen, 2018; Schimpfössl and Yablokov, 2017), I propose two initial parameters for disaggregating the concept of self-censorship. They accommodate the various kinds of intersecting heterogeneity of origins, mechanisms and effects that characterise the empirical forms of self-censorship. These parameters are the relationship between self-censorial practice and the social contexts in which it originates; and the interplay between free will, coercion and obligation.

To these ends, I present and analyse three illustrative accounts of self-censorship in contemporary Ukrainian journalism, as reported by journalists themselves. I draw these from my interviews with public broadcasting reporters and media watchdog professionals, which I collected in the course of a recent ethnographic research project investigating the meaning, value and practices of free speech in the context of the oligarchically dominated political economy of the media in Ukraine (Ryabinska, 2011, 2017). The accounts I have used are not representative. However, they give strong support to the argument that as scholars, we need not only distinguish between the different empirical forms of self-censorship which exist in the field but also to be reflexive about what makes one kind of self-censorship different from another. We need to put aside our preconceptions about the origins, and especially the logic of self-censorship, and explore the ways in which it occurs empirically within and across various contexts. My analysis of the three kinds of self-censorship among television journalists in Kyiv demonstrates the usefulness of such an approach. Later in this section, I briefly define these forms, as they emerged from the data. For now, however, I will consider the ways in which scholars have approached the socio-cultural variations in self-censorial practices.

The Encyclopedia of Censorship – a comprehensive catalogue of threats to freedom of speech all over the world – is a good place to start. On the 675 pages of its revised edition, ‘self-censorship’ is mentioned 66 times (Green and Karolides, 2005: passim). Yet the phenomenon is neither defined, nor considered worthy of an entry of its own. In the numerous country-focused articles that feature the term, the authors imply that it means the fearful anticipation of censorship or punishment, which in their turn result from authorities’ fear of free, critical speech (Green and Karolides, 2005: xviii). For example, this is the authors’ evaluation of the situation in turn-of-the-century Ukraine:

Self-censorship [in media] is commonplace in reaction to such pressures as control of access to affordable state-subsidized newsprint; dependence on political patrons who facilitate financial support from the State Press Support Fund; politically motivated visits from tax inspectors; and close scrutiny from government officials, especially at the local level. (Green and Karolides, 2005: 582)

One finds similar descriptions of self-censorship as a reaction to the perilous conditions of journalistic work (cf. Zeveleva, 2019) in other articles in the Encyclopedia which deal with countries beyond the apparently safe haven of Western democracies. These similarities reveal the authors’ analytical position – one from which self-censorship, no matter where, when and to whom it happens, always takes the same form and ‘spring[s] from the same source’ (Green and Karolides, 2005: xviii). Although this is not surprising, given the Encyclopedia’s political commitment to a similarly singular notion of freedom of speech, such conceptual universalism seems to be less a reflection of the actual self-censorial practice in contexts as diverse as, say, Ukraine, Colombia and Ghana, than a normative authorial imposition.

Elsewhere, within the small existing literature on self-censorship, one finds more analytical diversity. It is not difficult to find differing conceptualisations of self-censorship, for instance, as the internationalisation of socially sanctioned norms of behaviour (Schimpfössl and Yablokov, 2017), or as a voluntary, non-externally imposed constraint on speech that responds to the wider political context (Lee and Chan, 2009). This Special Issue is a case in point: individual contributors vary in their conceptualisations of self-censorship, its causes and mechanisms (e.g. compare the approaches of Zeveleva (2020) and Rožukalne (2020)). The differences stem from distinct, not entirely reconcilable, analytical positions on self-censorship taken by the authors; within each of the studies, however, self-censorship is understood as a phenomenon with a singular logic.

In this article, I draw on the work of scholars who have found it productive to identify different kinds of self-censorship within the framework of the same study. For example, Cook and Heilmann (2010: 3) distinguish between public and private self-censorship, depending on whether the subjects align their decision to censor themselves with an external, public censor. Extending Cook and Heilmann’s argument (2013), Jungblut and Hoxha (2017) go as far as to propose a matrix of 12 analytically possible forms of self-censorship ¹ which they apply, where appropriate, to the data gathered during their research on post-conflict journalism in the Balkans. The forms differ in accordance with their public/private motivation, their origins and the persons who might be affected by the failure to self-censor (Jungblut and Hoxha, 2017: 226–229). These axes of distinction provide a useful guide for empirical research. However, despite the authors’ recognition of various possible forms of self-censorship, they suggest that these operate according to a single (albeit somewhat flexible) logic: at the core of self-censorial practice is invariably an individual decision as to whether an utterance or a publication would be ‘in line’ with one’s own or peers’ professional judgement, one’s political ideology, superior’s opinions, market expectations and so on (Jungblut and Hoxha, 2017: 228). While individual conformity to expectations and pressures is potentially an efficient explanation for self-censorship, I believe that it is important not to settle the question of which social logic guides self-censorial practice in advance, without understanding actors’ own perspectives on what motivates their practices in particular circumstances (cf. Boyer, 2003 for a similar point regarding institutionalised pre-publication censorship). As I demonstrate below, Ukrainian journalists recognise that self-censorship can be motivated by a desire to honour personal or professional obligations as well as instrumental calculation of reciprocity; that it can happen automatically, and unbeknownst to oneself; and that conformity to social expectations is rarely a clear-cut matter from the actor’s point of view.

The partial explorations of the ‘repertoires’ of self-censorial practices, and the factors according to which scholars distinguishing one kind of practice from another, only highlight the fact that the literature on self-censorship has not undergone the kind of relativising shake-up that the ‘normative’ scholarship on censorship (see Darnton, 2014 on the normative-relative distinction) has been through since the 1980s, when a new approach to censorship began to emerge. Post-structuralist in their understandings of the locations and effects of power, scholars as diverse as Pierre Bourdieu (1991), Judith Butler (1997), Richard Burt (1994), Robert Post (1998) and Annabel Patterson (1984) ‘recast censorship from a negative, repressive force, concerned only with prohibiting, silencing, and erasing, to a productive force that creates new forms of discourse, new forms of communication and new genres of speech’. (Bunn, 2014: 53)

As Darnton (2014) suggests, underlying this shift was the relativisation and critique of the liberal commitment to seeing censorship as an enemy of freedom (see also Schimpfössl et al., 2020). The relativisation, however, happened in more than one dimension. If the earlier literature largely understood censorship as suppression of speech within top-down relations between censorial institutions and censored subjects (Bunn, 2015; Müller, 2004), the proponents of New Censorship Theory explored the dynamics of censorship in a much wider array of locales, identified a variety of forms of ‘regulation’ of expression and traced a greater diversity of causal relations determining censorship. In this way, New Censorship Theory has allowed for more nuance in understanding and explaining the empirical diversity of censorship, not least because identifying censorship as a constitutive factor of all expression made it possible to ask what allows subjects to see one kind of discursive constraint as different from another.

This article transposes these insights to the analysis of self-censorship among television journalists and media professionals in Kyiv. My interlocutors reported a range of different self-censorial practices, some of which overlap with one another, and in which there are some important differences that are not easily reducible to one dimension of distinction or one principle of differentiation. These practices and their interpretations constitute samotsenzura (Ukrainian and Russian for self-censorship) as a social fact among my interlocutors. Having analysed their accounts, and taken into consideration the fact that two research participants explicitly reflected on differences between various self-censorial practices, I divide the continuum of self-censorship they depicted into three categories. The names for these categories are mine: my respondents spoke about different self-censorial practices, but did not use separate names for them. The proposed three forms of self-censorship therefore result from my systematisation of the respondents’ accounts, made with a view to the most significant variations of reported practices.

The first form, pragmatic self-censorship, is locally understood to stem from journalists’ expectations of personal sanctions (positive or negative) within the social context of the workplace. Such self-censorship is a response to censorial injunctions or incentives, formulated by various powerful actors, and translated through concrete managerial hierarchies, editorial tasks and other relations that journalists navigate in their everyday work. The second form, ethical self-censorship, is described as deliberate and motivated by personal obligations to, or ethical concerns for, other people who could be affected by the journalist’s speech; its goal is the preservation and maintenance of social relations. The third form, affective self-censorship, arises involuntarily, as a lack of objectivity in reporting or an automatic repression resulting from personal sympathies, attachments and political positions, which might interfere with the public professional role of the journalist.

While pragmatic self-censorship is relatively well understood in the literature (e.g. Bourdieu, 1991), and ethical self-censorship overlaps with forms described by Jungblut and Hoxha (2017), it is significant that there is variation across practices in the three groups, and that they are reported within the same professional community, and indeed, in one case, by the same person.

In what follows, I detail the methods and the context of the study from which this article draws its materials, before moving on to discuss examples of the three forms of self-censorship. I close with a summary elaboration of the possible parameters for recognising the empirical heterogeneity of self-censorship.

Research methods and context

The cases presented in this article are drawn from materials collected over the course of 13 months of ethnographic research (between June 2017 and July 2018, and in January and March 2019) in Kyiv, Ukraine. Part of a collective project investigating freedom of speech as a lived value in four European contexts, the study aimed to explore the ideas and practices of freedom of speech, autonomy and professionalism among public service broadcasting journalists and media watchdogs after the Maidan Revolution of 2013–2014.

The study focused on Hromadske TV, an online multimedia news organisation created and run by reporters themselves. In addition, I researched a current affairs division within UA:PBC, also known as Suspilne, ² the Ukrainian public broadcasting company established in 2016 following a reform of the state-owned broadcaster. I conducted 60 biographical interviews in total with journalists, editors, producers and managers and did participant observation in the newsroom, at meetings and trainings, and following reporters on their beats. To better contextualise Hromadske and Suspilne, I interviewed people from the professional networks of my primary research participants within media watchdog, media development organisations, and mainstream privately owned TV channels. Taken together, the interviews present an informative picture of professional careers of journalists and media experts who mostly came to the profession in early- to mid-2010s and transitioned into public broadcasting as career option alternative to the work for private, ‘oligarchic’ media (Ryabinska, 2011).

The topics of self-censorship and censorship, though not directly included in the original research questions, emerged from the data. The interviews included questions about the respondents’ experience of external interference into their work, to which they frequently answered with stories about censorship. Many of them volunteered accounts of such interference without being directly asked, such as when I probed them on their reasons for leaving better-paid jobs at privately owned media organisations for relatively less secure work at Hromadske or Suspilne. Because I have studied public service broadcasting journalists and professionals at media development organisation funded by Western grants, my interlocutors might have been particularly keen to discuss (self-)censorship with me, as a lens through which to reflect on the meaning and limits of journalistic professionalism. In doing so, however, they gave accounts that have heuristic value in the absence of direct observations of self-censorial dynamics.

The examples selected from the interviews ³ comprise a variety of empirical situations. Some of these are my interlocutors’ reports of observed practice; others are based on their generalisations about their own or other people’s self-censorship. While they do not straightforwardly represent self-censorial practices, the three sets of examples illustrate the variation in what my respondents think self-censorship is in Ukraine. While one can reasonably suppose that the accounts of self-censorship reported and analysed here do not exhaust the range of possible forms of self-censorship, the differences among these examples do suggest some directions for a more nuanced scholarly understanding of this phenomenon in Ukraine and the broader region, and thus contribute to the programme of this special issue.

Pragmatic self-censorship

The main form of self-censorship reported by the journalists I interviewed can be termed pragmatic self-censorship. This is journalists’ socially structured subjective expectation of (dis)incentives for reporting within the social field they navigate. Pierre Bourdieu (1991; see also Zeveleva, 2020) called the force of such incentives, and the effect it has on linguistic production, ‘censorship of the field’. ‘[I]nherent in particular relations of linguistic production’ (Bourdieu, 1991: 79), such censorship does not need to be explicit and directly prohibitive. Because ‘[i]ndividuals implicitly and routinely modify their expressions in anticipation of their likely reception’, censorship stemming from the structure of particular fields ‘is transformed into self-censorship through the process of anticipation’ (Thompson, 1991: 19, original italics).

If all linguistic production is censored in virtue of being socially structured (the key insight of Bourdieu and other proponents of New Censorship Theory; see Schimpfössl et al., 2020), explicit censorial prohibitions or incentives are just a particular case of censorship which takes place in hierarchical organisational contexts. It is in this sense that my interviewees reflected on pragmatic self-censorship as a response to direct (although not necessarily unambiguous) commands, suggestions or hints. These are formulated by various powerful actors seeking to influence editorial agendas, and are mediated by managerial and editorial hierarchies, which makes editors key figures in both the maintenance of censorship within organisations, and in resistance to it (Koltsova, 2006).

One of the most reflexive accounts of self-censorship as an expectation of editorial pushback came from Nadia, an investigative reporter in her early 30s, who had earlier worked for one of the largest private TV companies. Before leaving due to a conflict with the management (on which more in the next section), Nadia had spent some 2 years in a newsroom that frequently produced stories about the war in Donbas and its effects on Ukrainian society. As our interview unfolded, Nadia spoke, without mentioning names or concrete situations, about the way that sensitive coverage of the war was handled in the newsroom. After the devastating defeat of the Ukrainian army in the battle of Ilovaisk in August 2014, there was a ‘complete prohibition’ in her newsroom on covering the aftermath of the battle; this ‘was a taboo topic for [President] Poroshenko’. When I asked how such a ‘taboo’ was formulated and translated into constraints on reporting, Nadia replied,

For example, [newsworthy] events [to do with the battle] were not being visited. ⁴ [. . .] Parliamentary committee meetings [where the defeat was discussed] were being ignored . . . [. . .] Originally, this was [a decision] at the level of the chief news editor. If [a manager] said that [covering] Ilovaisk was undesirable, no-one was trying to particularly push to broadcast the story. Journalists think, well, why would I go again to the editor and demand to make this story, if they won’t let me do it anyway. [. . .] With time, this turns into self-censorship.

The prohibition on covering Ilovaisk, Nadia explained, was never explicitly stated; it was perceived and interpreted as a pattern, and so came to be anticipated by journalists. As I interviewed more reporters with background in privately owned media, it emerged that the kind of self-censorship Nadia described was very common. Nadia and I spoke several months after a respected weekly Novoye Vremia (‘New Time’) published the results of a monitoring of the content of the flagship weekend news programmes of the four major TV channels ⁵ (including Nadia’s former employers) that together command a lion’s share of the Ukrainian television audience (Berdynskikh, 2017). The analysis found that only one in 20 news stories mentioning the president was critical of him, a disproportionately low percentage compared to the channels’ treatment of other officials and politicians. As media expert Yevheniya Kuznetsova was quoted as saying, ‘This means that either journalists engage in self-censorship, or there are unofficial arrangements [dogovorennosti] at the editorial level’. Nadia’s suggestion that both were the case, and that self-censorship responded to such ‘arrangements’, came as no surprise.

Saying that editorial discouragements ‘turn into’ self-censorship, Nadia does not specify whether they do so because journalists learn the pattern of editorial or managerial expectations within a particular newsroom, and so can stop self-censoring when such expectations change, or because they internalise, in a lasting manner, a more general feel for the censorial ‘conjuncture’. One interviewee, Oksana, who was employed by an internationally funded non-governmental organisation (NGO) seeking to reform the Ukrainian state’s relations with the media, suggested the latter,

Censorship, coming from the owner, the producer, the editor in chief, is one thing. But our [Ukrainian] journalists also have another ailment – self-censorship. They think, ‘If I’m working at a particular [private] media outlet, I can’t write about certain things a priori’. Or, ‘It would be good for me not to write about them’. [. . .] It’s way worse when someone who works for NewsOne or Espresso or 1+1 [private, ‘oligarchic’ media], then goes to work for an independent media, retaining this internal self-censorship. It’s somewhere here [points to her head] – it gets into one’s subconscious and it’s difficult to get out.

In her opinion, journalists internalise external censorial constraints not just in their content, but in principle – as an a priori expectation that one should format one’s expressions depending on an understanding of powerful interests that might be affected by one’s reporting. The mechanisms of self-censorship described by Nadia and Oksana could exist side-by-side, operating both through a generalised anticipation of the interests of the owner and concrete injunctions giving rise to self-censorship regarding particular topics or personalities. Katia, who at the time of our interview in 2017 was employed by a major media watchdog organisation in Kyiv, argued as much:

Everyone understands who is the owner of the media that they are working for . . . So when one starts working there, they understand: ‘Okay, so our owner is such-and-such, so, just in case, I won’t even try to write anything bad about his allies or anything good about his competitors. Because I understand what sort of media I’m working for’. This is self-censorship, which has a certain underpinning – either negative sanctions, or even, as I remember some journalist telling me, an internal editorial document stating: these people we cover like this, those – like that, still others we don’t mention at all, and yet other ones, only negatively.

Katia was quick to point out that more often than not ‘black lists’ originate from within the news organisation and are left unwritten, lest they be leaked to the public. She suggested that this practice differs from the system of unofficial censorship in Ukraine which existed in the early 2000s, when the administration of President Leonid Kuchma distributed temnyky (Ukr.; literally, ‘lists of topics’) to broadcasters: daily directives on what topics are to be reported and how, often with exact phrasing included (Dyczok, 2006). A contemporary report by Human Rights Watch provides a good description of the effects such explicit constraints had on journalistic self-censorship:

In the face of increasing pressures from editors to report in specific ways on a limited number of topics, journalists increasingly resort to self-censorship rather than face arguments with top editors, negative reactions from the presidential authorities, or the loss of their jobs and careers. [. . .] The majority of journalists choose to stay at their [TV] stations and either comply fully with strict editorial policy or seek ways to negotiate within the boundaries of the directives. [. . .] Some journalists reported that their self-censorship progressed to such a degree that they preferred to avoid discussing material at all rather than present only one biased side of a story or the disinformation requested in the temniki [Rus.]. (HRW, 2003: 31)

Temnyky represent an extreme case on the spectrum of explicitness of censorial command. As my interviews suggest, contemporary self-censorial anticipation among Ukrainian journalists develops in relation to rather more ambivalent directives or contextual cues that might be difficult to understand for outsiders. Maria, a former high-level editor at a large privately owned TV channel, suggested that desirable and prohibited topics, and the related editorial tasks, are communicated euphemistically: ‘People say, “it ought to be done,” or “but you understand . . . [that it must/must not be done]”’. When the former producer general of a small private TV channel read out to me his WhatsApp conversation with the channel’s CEO, I only picked up on the hints when he explained what he thought the conversation was about that. What at first seemed like a neutral conversation about a ‘film director’ [rezhisser] passing on his ‘request’ [pros’ba] to broadcast a ‘documentary that has been produced for him’ was in fact a veiled directive from the channel’s owner (the ‘director’) to slot into the channel’s programming an externally produced film targeting his political rivals. The extra information needed to understand the request had to be picked up from the larger context, and was thus by definition implicit. (My interlocutor claimed he had refused the request, and this contributed to his eventual decision to resign.)

The ambivalence of such requests, and rootedness in the shared social reality of a newsroom or a media organisation meant that my interlocutors needed to interpret such ambivalence appropriately. In a situation where a potential misreading of cues could incur a sanction (however mild), ambivalent censorship is all the more powerful for its vagueness and unpredictability, eliciting self-censorship that replaces the need for direct censorial control.

This, of course, means that actors’ interpretations and contextual knowledge may fail. Marta, formerly a news bulletin editor working under the direction of the abovementioned Maria at a large private TV channel owned by one of the richest people in the country, was fired after she broadcast a news story featuring a politician who was an enemy of the owner. She explained: ‘Surprisingly, [the owner] watched the news bulletin – he was furious, people said, and shouted he’d fire the entire output team’. She did not know about the animosity between the owner and the politician, even though her editor claimed she should have. In the interview, Marta effectively admitted that if she had anticipated that the story would elicit such a response, she would not have included it in the bulletin, or would have edited the politicians’ appearance out of it.

Marta’s failure to do this – a failure that mattered only because of a series of coincidences – resulted in a negative sanction. She made it clear to me that it was difficult to anticipate such a course of events, not least because the owner was unlikely to watch a midday news programme on a national holiday. Here, the censorial injunction seems to have been clearly formulated only after the event, through the owner’s reaction. If Marta did have a sense of needing to avoid broadcasting stories about personal enemies of the owner, she could not always be sure who these were; knowing the rules of the censorial game did not mean she would always know when to apply them.

Whereas some accounts of self-censorship in Ukraine (HRW, 2003; Ryabinska, 2017: 78) and elsewhere (Billiani, 2014; Jungblut and Hoxha, 2017; Lee and Chan, 2009) stress that it is committed out of fear of consequences, these examples suggest that when self-censorship takes the form of anticipation, this is not necessarily a fearful expectation of clear consequences of (not) censoring oneself, but rather a more-or-less vague weighing of multiple factors. Pragmatic self-censorship is a major form of journalistic self-censorship in Ukraine, whether it is the result of arrangements between politicians, media owners, managers and editors, as Nadia suggested; or of direct commands and lists of prohibited topics and personalities; or of implicit understandings of the ever-changing conjuncture of political or commercial interests and relations. It responds to, and is shaped by, the dynamics of power and control within particular organisations, which might be sufficiently predictable for it to develop into established ‘rules of the game’, as described by Schimpfössl and Yablokov (2017, 2020) and Zeveleva (2020), or remain more fluid, requiring a constant interpretation of contextual cues and situated expectations as to what should and should not be reported. This suggests that variations in the practices of pragmatic self-censorship correspond to particularities of hierarchical relations that journalists have to navigate within media organisations.

Ethical self-censorship

The second form of self-censorship I identify is based not on the expectation of sanctions, but on the maintenance of social ties. I call it ethical self-censorship to reflect my respondents’ understanding of it as an unambiguously deliberate, un-coerced practice that is motivated by an ethical concern for socially valued others, a sense of obligation arising from personal relations with them, or calculations of reciprocity in dyadic relationships.

In the interview, Nadia described her experience of leaving the large ‘oligarchic’ TV network where she had worked until 2016 as a reporter in a team producing a programme of investigations into ‘social’ issues – housing, healthcare, displacement and the human toll of the war in Donbas. She had previously worked as a political reporter, and told me she deliberately opted for a ‘social’ beat to avoid censorial constraints she associated with political news. Yet the ongoing war, and the government’s many failures to deal with its social and economic effects, had politicised Nadia’s new domain.

Before long, she explained, the channel’s management was attempting to interfere with her team’s critical coverage of the war. The team editor’s efforts to raise the issue with the management had no effect. Eventually, after coming into conflict with the channel’s CEO, the team decided to resign collectively. In the recent history of Ukrainian television and print journalism, there have been several prominent cases of resignations in protest at hostile takeovers (as with TVi and Forbes Ukraine in 2013, when collective resignations literally brought these media to a halt) or censorship (which led to a series of individual resignations at ZIK TV in the summer of 2017). In these cases, resignations were accompanied by statements pointing to the conflict and justifying journalists’ positions. I asked Nadia whether there was a similar statement in her case. She said,

I regret not having made a statement . . . [. . .] We had close relationships with many people at the channel, so we left silently. And no-one . . . Well, I thought it would be clear to everyone [in the journalistic community] – I thought it would be clear that when the entire team leaves, everyone would understand why we’d done that. [. . .] I regret that we did not state this [publicly]. [. . .] We had difficult relationships with the top manager of the channel who did not want to reach a compromise with us, but there were still many good journalists, good projects there . . . So, if we declared that we were leaving because of censorship, this would mean that there is high censorship at [the channel]. We understood that no-one would leave except for us, and that people somehow had to continue to work there. [. . .] I don’t know, maybe this would sound banal, but we did not want to undermine our colleagues. [. . .] Now I regret this.

While Nadia explains the decision to resign by reference to ‘difficult relationships with the top manager’, the team’s decision not to publicly denounce censorship at the channel is motivated neither by fear of reprisals from the management, nor by a preoccupation with personal consequences for those protesting the censorship. (The resignation was in itself a recognition that little could be changed about the situation; it was forced and thus a form of structural sanction for not accepting the management’s censorship and not resorting to self-censorial adaptation.) Rather, Nadia suggests, the team’s collective silence reflected their ‘close relationships’ with their colleagues and a concern for the reputation of those ‘good journalists’ who stayed on. An explicit denunciation of censorship at the channel would have implied that those who had not left were in accord with it. Although she does not call this an act of ‘self-censorship’, we can nevertheless recognise it as such, not least because Nadia herself came to think that she should have made a public denunciation. This suggests that while the journalists refused managerial censorship or the self-censorship which would have been necessary to conform to the management’s directive, they could still censor themselves in a different matter, where such silencing was seen as virtuous. Here is a case of self-censorship seemingly informed by an ethical concern about colleagues’ standing in the professional community.

Oksana, the NGO worker quoted above, suggested that deliberate self-censoring motivated by a desire to maintain certain social relationships might also happen in journalists’ relationships with their sources, in a social context different from that of relationships with colleagues or superiors in the newsroom. ‘One might [begin to] think, “I won’t be criticising this MP because he’s my friend, or because we drank together yesterday, or because tomorrow I’m going to ask him about a bill [in the Parliament]” . . .’ Pointing out several different forms of binding sociality – friendship, commensality, exchange of favours – Oksana foregrounds what in her view is a problematic calculation of self-interested reciprocity that ultimately leads to a confusion between the journalists’ impersonal professional duties (such as impartiality of reporting) with personal obligations arising from the particularistic relations they have cultivated in their personal or professional lives.

The two examples suggest that whether journalists find them to have moral value, as in Nadia’s case, or judge them as improper, as in Oksana’s, honouring such obligations to valued others might require deliberate acts of self-censorship. Such relationally motivated silencing works differently from pragmatic self-censorship. It does not entail a hierarchical relationship that organises sanctions and incentives, nor does it require expectation or internalisation of censorial prohibitions. In contrast, it appears to be deliberately calculated, motivated by the goal of maintaining social relationships that have intrinsic or instrumental value to the journalist. In their exercise of mapping out logically possible forms of self-censorship, Jungblut and Hoxha (2017) identify one which is similarly based on journalists’ consideration of the impact of their publications on others. Yet for these authors, such self-censorship remains an ‘individual self-restriction of one’s freedom of speech’ (Jungblut and Hoxha, 2017: 227). But if we allow that in such forms of self-censorship there might be more than one kind of freedom at stake, we can ask whether subjects might seek to restrict one kind of freedom (of speech) to realise or protect another. My respondents’ stories, for example, suggest a certain (limited) freedom of choice between the normative rationales (such as honouring personal obligations vs an impersonal professional code) governing one’s decisions in particular situations. In the last analysis, ethical self-censorship involves an exercise of subjective freedom within the social relationships one finds themselves in.

Affective self-censorship

The third form of self-censorship which emerges from my interviews is closely related to ethical self-censorship, yet differs from it in one important respect: it is involuntary. What I call affective self-censorship is understood by my respondents to arise from journalists’ personal sympathies and social attachments. It is a reflection of the ways in which social worlds inhabited by journalists shape their work, tacitly influencing what they might see fit to say, or indeed what they might say at all.

Anton, a senior political reporter at Hromadske with experience in privately owned print media, offered the following account when I asked how he deals with sources in positions of power:

My approach [as a journalist] entails some kind of detachment, but obviously I have certain sympathies of my own. I might for instance like some person, like what they’ve done, say, a particularly complex political move – I respect intelligent people. [. . .] But I understand that if I keep covering this person [all the time], self-censorship might turn on, and when something happens around this person, I might think: but he’s cool, we talk constantly, why would I spoil the relationship . . . and I won’t be objective.

Anton’s description of conditions for self-censorship is similar to the case of the relationally motivated silencing analysed above: ‘good relationships’ with political sources, he explained, might give rise to demands for favours and reciprocity, at which point the journalist might lose their impartiality. But there is something else at stake here. Note how Anton presents self-censorship as something almost automatic – a mechanism that ‘might turn on’ when triggered by subjective sympathies formed in the course of journalistic work. It is as if such self-censorship were inaccessible to the self-censoring subject themselves, rendering superfluous the question of whether or not such a form of self-censorship is coerced or deliberate. This is a kind of self-censorship that is more akin to psychological repression or bias than to fearful anticipation or deliberate withholding of speech.

In a similar fashion, Larissa, a programme editor at Hromadske, noted that while working as an investigative reporter at her previous job, she would abstain from reporting on anything to do with businesses that her family were involved in, knowing that she would ‘not be able to be objective’ about it. The strongest formulation of this form of self-censorship, however, came from Nadia. Speaking about war reporting, she suggested that this subject might present particular difficulties for journalistic objectivity because of the character of relationships one forms with one’s sources and subjects on the frontline:

Particularly at the beginning of the war, it happened so that typically in each newsroom the same people covered the ATO [the Anti-Terrorist Operation, a dated official euphemism for the Donbas war]. I have a friend working for [another channel]. He’s been in ATO since the first days [. . .] And . . . one forms very special . . . very humane relationships with the soldiers. And it’s a particular kind of atmosphere where you go there [to the front], it’s not like here in Kyiv where you have your peaceful life. There, people make bonds somewhat differently. And it so happens that these journalists, they cannot objectively cover the events there. And, well, I understand them in a way. And I myself . . . I understand that I myself would rather not produce stories about looting [by Ukrainian soldiers] because I have friends there [on the frontline], and my classmate died there. [. . .] In other words, I have my own attitude to this; I can’t, really. But I am convinced that this should be done; that there are people who can objectively report on this, and that this should be done.

From Nadia’s perspective as an outside observer, lack of objectivity in war reporting results in silence about facts that might be publicly relevant ⁶ and is synonymous with self-censorship. She empathises with her colleagues facing the difficult task of reporting in an objective – detached, balanced – way on people with whom they formed ‘very humane’ personal relations, and concludes that she herself would not be able to do so. Like Anton, though in her case more strongly, Nadia describes these relations as personal, that is, overflowing or perhaps preceding the professional persona of the journalist. Yet unlike Anton, she admits that one might simply be incapable of overcoming the affective impulse such personal relations introduce into one’s work.

These descriptions suggest certain similarities with ethical self-censorship – namely, the recognition of fact that journalists’ reporting is inevitably shaped in various ways not only by their professional duties and hierarchical relations in the newsroom, but also by social relationships they form on and off duty, which per se might have little to do with the substance of journalistic work affected by self-censorship. As described by my interlocutors, affective self-censorship reflects the ways in which individual journalists are shaped, as subjects, by their social ties over the long term.

Like the previous two forms of self-censorship discussed in this article, this form emerges out of the everyday social relations inhabited by journalists in their professional and personal capacities. Yet unlike the other two, which seem to lie squarely within the domain of journalistic work, this form of self-censorship emerges on the boundary between the personal and the professional. This distinguishes it from self-censorship as anticipation that inheres in the power dynamics of particular organisations, and from ethical self-censorship, which also appears situation-specific, even if guided by moral or instrumental judgements in the conduct of particular relationships. If pragmatic self-censorship represents a loss of professional autonomy due to hierarchically imposed and often punitive constraints, this involuntary self-censorship demonstrates how one’s autonomy as a journalist is always configured in relation to particular personal dispositions. Likewise, the involuntary, automatic character of this self-censorship, which we can glean from Larissa and Nadia’s admissions of being incapable of impartial reporting on particular topics, sets it apart from ethical self-censorship characterised by explicit deliberation.

Conclusion

Beginning with the common-sense understanding of self-censorship as a practice universally motivated by fear of punishment and persecution, I have explored three distinct forms of self-censorship emerging from my interviews with television journalists and media professionals in Ukraine. The general purpose has been to demonstrate that it is important and analytically productive to explore the empirical heterogeneity of self-censorship. More precisely, it has been to argue that the forms I identify and analyse in Ukraine operate in accordance with different if related logics (anticipation of sanctions, ethical regard for one’s obligations to others and socially shaped subjective bias). I have thus sought to contribute to the existing scholarship that distinguishes, analytically, between some forms of self-censorship (Cook and Heilmann, 2013; Jungblut and Hoxha, 2017), but is less flexible in recognising that these forms might significantly differ in their ‘mechanisms’, or that they can coexist within the same professional group, or the same individual, at a single historical moment.

Looking at different forms of self-censorship simultaneously, as phenomena occurring within the same regime of journalistic expression and sociality, allows us to ask what distinguishes these kinds of self-censorship from one another. So far, my analysis suggests that different social contexts, and different kinds of social relations navigated by journalists, give rise to different forms of self-censorial practice. If workplace hierarchies and collective dynamics structure reporting incentives, thus motivating individual conformity through pragmatic self-censorship, more equal dyadic relationships with colleagues and sources lead to self-censorship motivated by ethical concerns for others, the desire to honour obligations to them, or indeed by a calculation of favours to be reciprocated. As the examples of affective self-censorship suggest, the relevant social relations need not necessarily pertain to the professional domain. Therefore, the relationship between self-censorial practice and the concrete social dynamics in which it is manifested is one important parameter for empirically disaggregating the concept of self-censorship.

Another such parameter is the interplay of free will, coercion and obligation in self-censorship. Because self-censorship is so pervasively understood as a consequence of the external threats and oppression inimical to free expression (e.g. Jungblut and Hoxha, 2017; Ryabinska, 2017), it is often conceptualised as either an internal replication of an oppressive social dynamic, or an act of conformity. In either case, it becomes difficult to distinguish from censorship as such (see Schimpfössl et al, 2020); and while the descriptor ‘self-’ suggests a tension between personal agency and external constraint, by and large the self-censoring subject is taken to be a priori unfree (cf. Laidlaw, 2002). The findings emerging from my research complicate such understandings, suggesting that different forms of self-censorship vary in the degree of reflexivity and (self-)control that subjects have when censoring themselves. If pragmatic self-censorship is indeed a more or less voluntary, conscious response to external constraints and incentives as they are mediated through the everyday relations of power within media organisations, ethical self-censorship is not only deliberate, but also not constrained. The social logic of such self-censorship is one of ethical regard, care or expectation of reciprocity. In other words, it has to do with various forms of binding sociality that characterises relatively equal relationships and cannot be described as coercive in the same sense as censorial commands that one learns to anticipate in the newsroom. Finally, self-censorship as subjective bias appears to be neither completely deliberate, nor conscious, nor coerced, but automatic in the sense of psychological automatism.

These dimensions of difference are impossible to distil into a matrix of logically possible universal types of self-censorship, such as Jungblut and Hoxha (2017) have outlined. The forms of practice that these dimensions characterise and distinguish from one another in the cases discussed in this article are situated in the particular socio-historical context of post-Maidan, war-time Ukraine. They are not generalisable per se, beyond the analytical point about the dimensions (parameters) of heterogeneity that might be productively investigated.

More research is needed to understand how, when and why the same journalists engage in different forms of self-censorial practices; how such practices relate to each other and to other forms of cultural regulation of speech (Mazzarella and Kaur, 2009); how they are situated ‘within the communicative paradigms of specific historical contexts’ (Bunn, 2015: 28); and what specific social arrangements make them possible. It is likely that answers to these questions will be highly specific to the socio-historical contexts in which they are posed. This article and the Special Issue as a whole make a step in this direction.