Justifying Self-censorship : A Perspective from Ethiopia

This study investigates self-censorship practices in Ethiopian state media institutions. Through indepth interviews with 61 journalists, the study discloses extensive use of self-censorship on the part of journalists who try to conform to the expected reporting style of the state media. The journalists are largely critical of self-censorship, but continue with the practice despite their reservations. The study suggests that editors and reporters assume a set of underlying justifications to validate the practice on a personal level and make it appear professional for outsiders. The justifications are found to follow four lines of argument: (1) relegation of ethical responsibility; (2) elasticity of journalistic editing; (3) confidence in critical audiences; and (4) adherence to social responsibility. It is further found that there is a remarkable discrepancy between the relatively open-minded official editorial policy of the Ethiopian state media and the restrictive reporting practices followed by the journalists. It is suggested that discourses of fear play a significant role in the reproduction of self-censorship in the concerned media organizations. The phenomenon of self-censorship is widespread in African media practices. It is found in both the private and the state-owned media; in new and old media; during times of tension and indeed also in more peaceful periods. It is, according to Derek Forbes, as common as political influence on the media, thus he suggests that ‘political interference and self-censorship go together in African journalism’ (2005, 61). Even so, scholarly analyses of the practices are limited. To the extent that self-censorship is examined, the analyses tend to focus on the political and legal conditions that cause self-censorship instead of the journalists’ own experience of the practice and how they reflect on it. This study explores the felt conflict between self-censorship and journalistic professionalism. Through individual interviews with 61 reporters and editors in the Ethiopian state media, I try to identify the boundaries between acceptable and unacceptable self-censorship practices as experienced by the journalists, and to unearth the journalists’ rationalization for putting restraints on themselves in ______________________________ Westminster Papers in Communication and Culture © 2010 (University of Westminster, London), Vol. 7(2): 98-121. ISSN 1744-6708 (Print); 1744-6716 (Online) Skjerdal, Justifying Self-censorship... 99 professional reporting. I shall pinpoint four justifications for self-censorship that emanate from the accounts of the informants: (1) relegation of ethical responsibility; (2) elasticity of journalistic editing; (3) confidence in critical audiences; and (4) adherence to social responsibility. Although the subject of study is media practices in Ethiopia, I believe that the findings are relevant to other media contexts as well – especially in Africa, where the media typically operate under similar conditions of government interference and unstable political circumstances. That the findings have comparative value is substantiated by the fairly universal features of self-censorship practices observed across the African continent, although distinct political climates and diverse media situations obviously cause local variations. Ethiopia is similar to many other African media societies in that diversity in ownership only came with multi-party democracy in the 1990s, although the authorities proved to be much more reluctant to open up for private actors in broadcasting than in the print sector. A majority of the journalists still work for the state media institutions where they feel obliged to report in line with the political programme of the government of the day. Defining self-censorship in media practice is not necessarily a straightforward issue. It stretches from a wide understanding, seeing self-censorship as an everyday practice for any journalist anywhere in the world, caused by the inevitable selection and de-selection processes while reporting and editing; to a narrow definition, entailing only those practices which are performed for the sake of excluding information from publicity due to felt threats by public authorities. The understanding used in this article comes closer to the latter than the former, as I define self-censorship as the withholding of journalistic material due to felt external pressure. Importantly, this definition involves not only limitations caused by government interference, but also those provoked by other actors and conditions, including cultural expectations and in-house procedures. The concentration is thus largely similar to that of several other studies which have investigated journalistic self-censorship, such as Chin-Chuan Lee’s research into journalism cultures in Hong Kong, in which he defined self-censorship as: a set of editorial actions ranging from omission, dilution, distortion, and change of emphasis to choice of rhetorical devices by journalists, their organizations, and even the entire media community in anticipation of currying reward and avoiding punishment from the power structure. (1998, 57) The focus here too remains on unwanted interference, pertinently distinguished from intentional routine editing which takes place in regular journalistic production. Self-censorship is difficult to measure. Even if it is agreed that this type of censorship is a key characteristic of oppressive media environments, press freedom groups are left with the challenge of how to quantify the activity in order Westminster Papers in Communication and Culture 7(2) 100 to make it comparable between countries and media organizations. As such it stands in contrast to more easily identifiable indicators like incidents of official censorship, imprisonments of media workers and the issuing of media legislation, which more frequently form the ground for comparison across countries for media advocacy groups (and in scholarly analysis too, e.g. Berger, 2007). Press freedom groups habitually refer to self-censorship as one of the key indicators in their yearly country reports (for instance those of Freedom House, Reporters sans frontières and International Research and Exchanges Board/IREX), but, as pointed out by Shelton A. Gunaratne, when blended with indicators of other types of repressive actions, ‘its usefulness is debatable’ (2002, 365). What adds to the challenge is that the experience of self-censorship is individual, even within the same newsroom. Thus one radio journalist from Rwanda imparts that ‘there was no question of selfcensorship’ in media practice in the local radio stations (‘Radio contributes ...’, 2007, 30), while researchers who have investigated radio reporting in the same country conclude that ‘most journalists admit to self-censorship’ (Myers, 2008). This illustrates that the experience of self-censorship is diverse and personal, and that different actors may not read the same meaning into the term when they refer to ‘self-censorship’. It also accentuates that it is a phenomenon that calls for qualitative in-depth research to be understood and assessed in different media cultures. Self-censorship in the African Media As several studies suggest, there is some sort of relationship between selfcensorship and the overall political climate in the country (Amin, 2002; Nadadur, 2007; Simons and Strovsky, 2006). The political environment also appears to have an impact on the journalists’ perception of self-censorship. For example, a comparative study between journalists in China, Hong Kong and Taiwan found that journalists in China were least opposed to censoring themselves, perhaps stemming from the long tradition of state-controlled media in the country (Lo et al., 2005). This may resonate with experiences from Africa, although research in the field tends to be country-focused rather than comparative. The following section thus attempts to give an introduction to scholarly work about selfcensorship in Africa in three focus areas: the phenomenon’s relation to the political environment; differences in practices between the state media and the private media; and the correlation between self-censorship and national media legislation. I will also briefly discuss how factors such as commercial pressure, cultural expectations and lack of professionalism may have an effect on journalistic self-censorship, as suggested by various African researchers. In Africa, as elsewhere, self-censorship is usually associated with societies that have strong tendencies to exert state control over the media, such as Eritrea (Aaron, 2006), Rwanda (Thompson, 2007a) and Zimbabwe (Mukundu, 2006). The rationale is that severe media regulation is a characteristic of defensive states Skjerdal, Justifying Self-censorship... 101 willing to employ a range of means to control their citizens. The outcome is a high degree of fear for authorities, and media institutions are not exempt. However, self-censorship is observed in more open media societies too, such as in Kenya (Barland, 2005; Ongong’a, 2010) and Ghana (Kafewo, 2006). Even the most liberal media culture on the African continent, that of South Africa, has been critiqued for fostering self-censorship in the newsrooms. The critique in this regard is particularly aimed at the public broadcaster SABC (South African Broadcasting Corporation), which has been accused of being a tool for the ruling party ANC (African National Congress). Two cases may be mentioned; first, a critical documentary about the then President Thabo Mbeki which was removed from the programme schedule of SABC in May 2006; and, second, the blacklisting of government-critical commentators by the managing director of News and Current Affairs in SABC later that year (FXI, 2008; IREX, 2008; Teer-Tomaselli, 2008). A commission of inquiry found the managing director guilty of breaking fundamental journalistic principles when ousting the 

professional reporting.I shall pinpoint four justifications for self-censorship that emanate from the accounts of the informants: (1) relegation of ethical responsibility; (2) elasticity of journalistic editing; (3) confidence in critical audiences; and (4) adherence to social responsibility.Although the subject of study is media practices in Ethiopia, I believe that the findings are relevant to other media contexts as well -especially in Africa, where the media typically operate under similar conditions of government interference and unstable political circumstances.That the findings have comparative value is substantiated by the fairly universal features of self-censorship practices observed across the African continent, although distinct political climates and diverse media situations obviously cause local variations.Ethiopia is similar to many other African media societies in that diversity in ownership only came with multi-party democracy in the 1990s, although the authorities proved to be much more reluctant to open up for private actors in broadcasting than in the print sector.A majority of the journalists still work for the state media institutions where they feel obliged to report in line with the political programme of the government of the day.
Defining self-censorship in media practice is not necessarily a straightforward issue.It stretches from a wide understanding, seeing self-censorship as an everyday practice for any journalist anywhere in the world, caused by the inevitable selection and de-selection processes while reporting and editing; to a narrow definition, entailing only those practices which are performed for the sake of excluding information from publicity due to felt threats by public authorities.The understanding used in this article comes closer to the latter than the former, as I define self-censorship as the withholding of journalistic material due to felt external pressure.Importantly, this definition involves not only limitations caused by government interference, but also those provoked by other actors and conditions, including cultural expectations and in-house procedures.The concentration is thus largely similar to that of several other studies which have investigated journalistic self-censorship, such as Chin-Chuan Lee's research into journalism cultures in Hong Kong, in which he defined self-censorship as: a set of editorial actions ranging from omission, dilution, distortion, and change of emphasis to choice of rhetorical devices by journalists, their organizations, and even the entire media community in anticipation of currying reward and avoiding punishment from the power structure.(1998,57) The focus here too remains on unwanted interference, pertinently distinguished from intentional routine editing which takes place in regular journalistic production.Self-censorship is difficult to measure.Even if it is agreed that this type of censorship is a key characteristic of oppressive media environments, press freedom groups are left with the challenge of how to quantify the activity in order to make it comparable between countries and media organizations.As such it stands in contrast to more easily identifiable indicators like incidents of official censorship, imprisonments of media workers and the issuing of media legislation, which more frequently form the ground for comparison across countries for media advocacy groups (and in scholarly analysis too, e.g.Berger, 2007).Press freedom groups habitually refer to self-censorship as one of the key indicators in their yearly country reports (for instance those of Freedom House, Reporters sans frontières and International Research and Exchanges Board/IREX), but, as pointed out by Shelton A. Gunaratne, when blended with indicators of other types of repressive actions, 'its usefulness is debatable ' (2002, 365).What adds to the challenge is that the experience of self-censorship is individual, even within the same newsroom.Thus one radio journalist from Rwanda imparts that 'there was no question of selfcensorship' in media practice in the local radio stations ('Radio contributes … ', 2007, 30), while researchers who have investigated radio reporting in the same country conclude that 'most journalists admit to self-censorship' (Myers, 2008).This illustrates that the experience of self-censorship is diverse and personal, and that different actors may not read the same meaning into the term when they refer to 'self-censorship'.It also accentuates that it is a phenomenon that calls for qualitative in-depth research to be understood and assessed in different media cultures.

Self-censorship in the African Media
As several studies suggest, there is some sort of relationship between selfcensorship and the overall political climate in the country (Amin, 2002;Nadadur, 2007;Simons and Strovsky, 2006).The political environment also appears to have an impact on the journalists' perception of self-censorship.For example, a comparative study between journalists in China, Hong Kong and Taiwan found that journalists in China were least opposed to censoring themselves, perhaps stemming from the long tradition of state-controlled media in the country (Lo et al., 2005).This may resonate with experiences from Africa, although research in the field tends to be country-focused rather than comparative.The following section thus attempts to give an introduction to scholarly work about selfcensorship in Africa in three focus areas: the phenomenon's relation to the political environment; differences in practices between the state media and the private media; and the correlation between self-censorship and national media legislation.I will also briefly discuss how factors such as commercial pressure, cultural expectations and lack of professionalism may have an effect on journalistic self-censorship, as suggested by various African researchers.
In Africa, as elsewhere, self-censorship is usually associated with societies that have strong tendencies to exert state control over the media, such as Eritrea (Aaron, 2006), Rwanda (Thompson, 2007a) and Zimbabwe (Mukundu, 2006).The rationale is that severe media regulation is a characteristic of defensive states willing to employ a range of means to control their citizens.The outcome is a high degree of fear for authorities, and media institutions are not exempt.However, self-censorship is observed in more open media societies too, such as in Kenya (Barland, 2005;Ongong'a, 2010) and Ghana (Kafewo, 2006).Even the most liberal media culture on the African continent, that of South Africa, has been critiqued for fostering self-censorship in the newsrooms.The critique in this regard is particularly aimed at the public broadcaster SABC (South African Broadcasting Corporation), which has been accused of being a tool for the ruling party ANC (African National Congress).Two cases may be mentioned; first, a critical documentary about the then President Thabo Mbeki which was removed from the programme schedule of SABC in May 2006; and, second, the blacklisting of government-critical commentators by the managing director of News and Current Affairs in SABC later that year (FXI, 2008;IREX, 2008;Teer-Tomaselli, 2008).A commission of inquiry found the managing director guilty of breaking fundamental journalistic principles when ousting the commentators.However, the fundamental difference between the alleged self-censorship practice in SABC newsrooms and self-censorship in oppressive societies is that the South African media environment encourages the different media outlets to scrutinize each other, thus fulfilling their obligation as self-regulatory entities.It follows that selfcensorship can hardly be eradicated altogether in any given media society; as long as there are external parties which have an interest in what is covered in the media, the chance of self-censorship is immanent.This emphasizes the need to differentiate between degrees of self-censorship in various societies and subsequently inquire whether it is either systematic or more arbitrary by nature.
As is evident from the above, self-censorship is particularly associated with the state media and official institutions (cf.Lo et al., 2005).This is perhaps only to be expected given that official outlets are customarily framed as mouthpieces for the authorities rather than public service media channels proper.Even if there has been a general drive towards more independent journalism in the African media (Blankson, 2007;Tettey, 2009), most publicly owned media institutions on the continent continue to nurture relationships with the authorities; sometimes in their organizational structures, but always in the minds of their audience.The state media carry on in a symbiotic relationship with the authorities whereby journalists are given easier access to services and information in return for friendly coverage of the government.Ethiopia is one example, where until recently only the state media had access to government press conferences.In Mozambique, another country where the state still controls a considerable portion of the media, journalists claim that the government has actively promoted self-censorship through the Higher Council of Social Communication (Feustel et al., 2005).The inducement to self-censorship tends to be covert rather than in the open, however, as governments do not want to be seen as enemies of press freedom through blatant dispersion of fear in the media industry.
On the other hand, governments are not always successful in stimulating selfcensorship.Assessing the state media in Cameroon in the 1980s, Francis B. Nyamnjoh, Francis Wete and Tangie Fonchingong describe how: 'despite the political socialization and repressiveness of the civil service, not many journalists did conform [to the government's communication strategy]; not everyone employed self-censorship to a level satisfactory to the authorities' (1996,52).The reason for the 'disobedience' was first and foremost a felt conflict with values of professional journalism that they had -ironically -learnt at state-run journalism training, according to the researchers.Expectations of professional and independent reporting by the public also played a role.Today, there appears to be increased conviction regarding the importance of independent reporting among public officials in Cameroon as well, but the public's perception of the state media outlets remains one of loyal support for the ruling power (Alobwede, 2006).
The history of self-censorship in the African media has a certain relationship with the continent's overall political history.After independence, it became important for the newborn governments to use the media to promote national cohesion and economic growth.The outcome, inevitably, was state-run media outlets which had clearly politically motivated editorial policies.The Nigerian government's takeover of the independent media from 1964 to 1974, for instance, unavoidably marked the introduction of self-censorship in Nigerian journalism, according to Onuma O. Oreh (1976).Likewise, it has been suggested that the nationalization of key media outlets led to state control and accompanying self-censorship in for instance the Egyptian (Ismail, 2008) and Tanzanian (Sturmer, 1998) Self-censorship has not only been associated with the state media, however.The private media are often reported to have put constraints on themselves for fear of legal or political reactions as well.However, while self-censorship is always an issue with state-operated media, it tends to be an issue with the private media largely only to the extent that there exists repressive media regulation.Self-censorship in the private media is therefore categorical in Zimbabwe (Dube and Gonçalves, 2004;Mukundu, 2006), while it is conditional in Kenya (Barland, 2005;Kalyango, 2008).
The relationship between repressive media laws and self-censorship is commonly assumed, and two sets of laws are particularly in focus: defamation or libel laws, and laws concerning national security, sedition or treason.The maximum penalty for transgressing the latter tends to be harsh, in worst cases the death penalty; and potential crimes are a prioritized area for many legal authorities.With regard to defamation or libel, private stakeholders may be as threatening as public authorities, although the prosecution of reporters and editors by civil society agents sometimes concurs with the interests of the authorities.In Angola in 2003, an editor of a private newspaper was summoned by the police after disclosing alleged links between a private charity headed by Angola's president and Brazilian drug dealers.The editor was allowed to leave after questioning, but subsequently faced 10 charges of defamation from individuals in the charity organization (Marques and Pearce, 2004).The fear of criminal charges is therefore a major impediment to journalistic activity in heavily media-regulated societies, and cultivates self-censorship.This is illustrated by a survey of Ugandan journalists conducted by Peter Mwesige, in which he found that more than half, 53 percent, cited official laws such as sedition and libel as a 'major limit on journalistic freedom ' (2004, 80).
The case of Uganda further explicates the troublesome relationship between legal measures and self-censorship in that, even if the penal code contains serious penalties for defamation or seditious material (Khamalwa, 2006), the fact of the matter is that few persons have been convicted under these clauses (Walulya, 2008).Nevertheless, the laws have a chilling effect, for instance in the reporting of corruption, because the journalistic investigation process may involve government officials and other public figures who have the resources to take legal action.For journalists and media owners, however, it is a pertinent question whether selfcensorship should always be an obvious reaction to legal restrictions.Indeed, after having talked to numerous reporters and editors in the African media, Helge Rønning (2005) suggests that the general perception is that there are far more legal restrictions than is actually the case.This may also stem from the phenomenon that media practitioners sometimes find it convenient to blame repressive conditions -because doing 'protocol journalism' (i.e.reporting official information issued by the state) involves far less effort than for instance investigative reporting.
With regard to censorship issues in Uganda, it is also pertinent to mention that the country, like a number of other African nations, has recently introduced an Anti-Terrorism Act which ultimately makes it a capital offence to pass on information that can aid terrorist actions.The act seriously increases self-censorship among journalists covering the conflict-ridden area of Northern Uganda, argues Linda Nassanga Goretti (2007;cf. McNamara, 2009).This is yet another indication that the decline of strong state media organizations does not automatically mark the end of self-censorship, because self-imposed censorship exceeds organizational models and ownership structures.
Beyond issues where political interests are directly involved, self-censorship is also induced by commercial pressure and cultural expectations.Advertisers do not only represent a potential threat to independent journalism in affluent media markets in Europe and North America (Craig, 2004), but also in for instance Uganda (Khamalwa, 2006) and Kenya, where 'negative stories are not run in order to protect revenue' (IREX, 2008, 276).In Nigeria, the oil industry is a major incentive for journalistic self-censorship -probably more so than unfortunate media laws, according to Anya Schiffrin (2009).As regards cultural expectations, journalists are challenged by traditional limitations, for instance in health reporting (Finlay, 2004;Mano, 2004).Self-censorship is also a common working method for journalists when covering gender (Tom, 2008) and religion (Banda, 2003;Servant, 2003).Last, and probably most significantly, self-censorship is instinctively applied in any coverage of ethnic conflict.It would suffice here to mention Rwanda, where ethnicity is considered a no-go area for the media (Waldorf, 2007), but it remains a sensitive area for reporting in virtually every African country.'You won't find a single journalist who will report freely on ethnic or religious issues', claims an editor from a private newspaper in Ethiopia (personal interview, 5 May 2009).Ethnic issues, besides issues where national security is involved, appear to be one of the most quoted reasons for self-censorship among Ethiopian journalists -both in the private and state media.
Self-censorship is rightly associated with journalistic unprofessionalism (Sunday, 2007) and poor training (Rønning, 2005), though the practice might in some cases be ethically defended as the lesser of two evils.Self-imposed censorship has indeed made it possible to resume journalistic practices in unstable societies, even to the extent that it is essential for survival, such as in Somalia (IREX, 2009).Selfcensorship is therefore not only a working method found in unprofessional media outlets, but is indeed part of the record of any significant media organization in Africa.Foreign correspondents of media outlets with an international reputation also admit having utilized self-censorship in order to subsist in the African environment, as explained by veteran journalist and editor Xan Smiley of Africa Confidential: Heavy self-censorship is a standard feature of reporting by Western journalists everywhere in the Third World, but nowhere is it more rigorously in force than in black Africa.For one thing, a reporter has to be as kind to the country he is writing as possible, if he is to have any hope of returning.(Smiley, 1982, 70, quoted in Terrell, 1989, 150) In the same way that self-censorship practices do not always indicate lack of professionalism, the presence of self-censorship doesn't necessarily mean absence of critical journalism.Discussing the democratic potential of the East African media, Jens Barland concludes that 'despite self-censorship, media's activities in Kenya have shown that a free and vibrant press can be useful in effecting transparency, democracy and good governance ' (2005, 88).Conversely, in Morocco, the decline of government-imposed self-censorship among journalists due to the increasingly less predominant role of the state media and the growth of private media outlets has not meant a blooming of quality journalism, mainly due to the fact that the new outlets have acquired a new form of partisanship (Douai, 2009).These examples serve to demonstrate that the correlation between selfcensorship and journalistic quality is unsystematic, even though experiences of heavy self-censorship are commonly associated with repressive media societies.

The Ethiopian Media Situation
Ethiopia is characterized by strong state media institutions, especially in the broadcasting sector.The only television company is state-run (Ethiopian Television, ETV), and nationwide radio is operated by the state as well (Ethiopian Radio).The government has recently opened up the airwaves for private radio stations, and four licences for commercial FM stations have been granted since 2006.As regards the print media, some 33 private newspapers are on the market in addition to four state-run ones, most of which are weeklies. 1The newspapers are mainly distributed in the capital city of Addis Ababa, which accommodates less than 5 percent of the approximately 80 million population in the country.
From 1974 to 1991, the country was ruled by the Derg regime with extensive use of military means and repression of fundamental human rights.The downfall of the regime in 1991 led to great enthusiasm in the emerging media market and a sudden growth in newspaper titles and magazines.However, many publications were highly sensational and remained in the market only for a short period of time (Shimelis, 2002).Throughout the 1990s, the new EPRDF (Ethiopian People's Revolutionary Democratic Front) government began sanctioning the private media and imprisonments of media workers became commonplace, especially during election periods.The latest grave crackdown on the private media occurred after the national elections in 2005 when 14 journalists and editors were put in custody for disrupting public order, only to be found not guilty by the court after more than 1.5 years in prison.Despite irregular reactions from authorities, the private media outlets have managed to carry critical journalism -albeit to varying degrees (Skjerdal and Hallelujah, 2009;Wondwosen, 2009).
The state-operated media, for their part, have continued to operate under the protection of their owner and have had certain privileges over the private media, like easier access to information and better opportunities for journalism training.Due to the size of the national broadcasting institutions, the state media cater for the largest contingent of journalists in the country. 2 An interesting feature of Ethiopian journalism, which sets it apart from most other journalistic cultures on the continent, is the overwhelming use of languages unique to the country (particularly Amharic and Oromiffa).This contributes to preserving the local culture through the media, but at the same time makes Ethiopian media content largely inaccessible to outsiders, thereby possibly preventing criticism from the international community (Skjerdal, 2009).As regards the content in the state media outlets, much of the material is presented as protocol news; in other words, it has the form of official information emanating from the public administration.Investigative journalism has not been much featured in Ethiopian state media channels.
Researchers point to self-censorship both in the private and state media in Ethiopia (Abebe, 2007;IREX, 2009), but the practice appears to be more present in state media newsrooms (Gezahegn, 2009;Nebiyu, 2009;Skjerdal, 2008).

Methodology
The present study is primarily based on in-depth interviews with reporters and editors in three Ethiopian state media institutions.A total of 61 journalists were interviewed, mainly from Ethiopian Television (ETV), The Ethiopian Herald (English daily) and the Ethiopian News Agency (ENA).The interviews were conducted between November 2007 and May 2010.Because of the sensitive nature of the research topic, most interviews took place outside the worksite.An exception is The Ethiopian Herald, which is less staffed by political appointees and has a more relaxed atmosphere than the broadcaster ETV and the news agency ENA.A number of the interview sessions in The Ethiopian Herald therefore took place within the newspaper's own offices.In addition to personal interviews, unstructured newsroom observation has been used to some extent in the study.
Efforts have been made to ensure open conversation.In most cases, I felt that the informants were open and willing to speak their mind, even if it meant criticizing their media institution and colleagues as well.In a few cases, probably less than five, I sensed that the interviewees were defensive; either in order to protect their owner or for fear that I would abuse the information they gave.
I rarely uttered the word 'self-censorship' in the interviews unless the informants used it in the first place.This strategy was used in order to prevent any preconceptions of the term; any direct questioning of 'self-censorship' could for instance easily divert into an argument over official censorship, which is indeed a rare phenomenon in Ethiopian media operations.I have rather tried to let the journalists talk about the daily news production in their own words and asked them to explain what type of challenges and constraints they face in their day-today work.
The private media were not included in the study because their modus operandi is different and their use of self-censorship takes a different form than that of the state-owned media, as seen in the review of African media practices above.Nevertheless, a similar study into the private media could prove interesting in order to decide the actual levels of independence of these media organizations and to further grasp the nature of journalistic professionalism in news media outlets which are often a result of discontent with the state media.
The research questions are divided into one of a descriptive and one of an analytical nature: 1. What characterizes self-censorship practices in the Ethiopian state media?2. How do journalists justify the practices -professionally and ethically?

Adapting Self-censorship Practices
The great majority of reporters in the Ethiopian state media -like those in the private media -have no degree or diploma in journalism (Birhanu, 2006).They often possess an undergraduate degree in some other area, in many cases language studies and literature,3 but professional journalism is learnt in the form of in-house short-term training, or, more importantly, on the spot as they begin practising as junior reporters.These processes are important because they include not only the learning of professional journalistic principles, but also an adaptation to the journalistic culture of the institution.The training varies between the different state media departments from no formalized introduction at all to an apprentice-like initiation period lasting several months.
A senior journalist in ETV recalls how he was introduced to journalism when he first began working there: On the first day in the station, they don't even try to let you know how ETV works.You just learn from seniors, from colleagues.They attach you to a senior reporter and you'll go with him for a week.The coming week you go alone.[...] When you do something wrong, something that is against the editorial policy, only then do you read the editorial policy.(ETV journalist, personal interview, 23 November 2007) This resembles the initiation period that several reporters have had in The Ethiopian Herald.The teaching philosophy consists of adapting a journalistic style which has emerged after many years of conveying official news.The key to proper reporting is to be found within the organization itself: When you start, you will be given stories to read.For some weeks, months, they simply tell you to read.It gives you an idea on how stories are written.You follow their steps.(Editor in The Ethiopian Herald, personal interview, 17 November 2007) An essential part of the adaptation process is learning the expectations of superiors and getting to know the style of official reporting.It begins with copying the standard formatting and writing style of bulletins and news articles.A set (though unwritten) formula is to be used when compiling a news story, and the first sentence of the story typically contains a quote from a government official who makes a statement which is either positive news for the country or demonstrates the thoughtfulness of the person.A standard opening of an article in The Ethiopian Herald is exemplified by the following front-page story which concerns a respected university professor who has just received a distinction: 'President Girma Wolde-Giorgis said that the outstanding achievement Prof. Gebisa Ejeta has made is immensely contributing to the national effort to reducing poverty and ensuring food security' (Mengesha, 2009).In typical manner, the focus is on the development efforts of the honouree; it is inconceivable that the state-run newspaper would pay attention to any critical remkarks about the government that the professor may have expressed in his scholarly writing.
To report properly for the state media, the journalist also needs to know who is allowed to speak.A state representative, often a cabinet minister, is essential to any news story.However, the editorial policy of the Ethiopian state media prescribes balance, and opposing views are supposed to be expressed if there is an obvious conflict in the subject matter (Ethiopian News Agency, 2003;Ethiopian Press Agency, 2003;Ethiopian Radio and Television Agency, 2005).It is nevertheless self-evident that the source who is allowed to speak on behalf the government ought to be communicated as the official and 'neutral' point of view, even if the reporter would prefer to present the view on an equal basis with potentially opposing views.The various sources are also duly attributed with titles, honours and -if there is an interview entailing direct questions from the reporter -polite salutations (e.g.'Your Excellency').
Routine procedures are vital to the functioning of state media news organizations.
A typical day for a reporter begins with a briefing by the immediate editor in which news items are assigned to the various reporters, followed by a few hours when the reporters goes out to collect information, after which the story is written and edited in the afternoon.Most news items are predetermined, for instance emerging from a press conference concerning the signing of a development project agreement.Sudden events rarely get any substantial coverage.The journalistic work in the state media is therefore not very exciting, maintains a journalist who has worked as editor for both the private and the state media.A jemari (beginning) reporter in ENA explains that there is no point for him to stay long when he goes out to cover an official event: What I need very much is the speech by the Minister or any other government official.From this speech I pick my lead.This is the way the government media cover events.Most of us will leave after the opening session.(Reporter in ENA, personal interview, 17 November 2009) After the speech is over, he asks for a short interview with the Minister, who will usually be kind to take the time to respond to a few questions.The reporter in this case says he feels free to ask critical questions too, but believes it is a waste of time because he would not get such material published since 'my chief may not accept it'.Asked whether the editors customarily alter the contents of his articles, the reporter says that they did that when he first joined ENA, but hardly any more.'I've learnt.I know what kind of news they don't change.Usually the editor will only be a copy editor' (reporter in ENA, personal interview, 17 November 2009).On rare occasions, the reporter, who has been with ENA for 2.5 years, has experienced that the editors have discarded his story altogether.This was for example the case when he was sent off to cover an event at the United Nations Conference Centre in Addis Ababa where a new UNCTAD (United Nations Conference on Trade and Development) report about the Least Developed Countries (LDC) was presented.Ethiopia came out badly in the report, a fact that was duly reflected in the final article compiled by the jemari.The article was subsequently dumped by the editors in the news agency.'It was such a big story, and it was not covered by ENA,' the reporter recalls with regret.
For this particular journalist, the feeling of compromised journalistic standards in the state media is sincere.He is among the first graduates with a BA degree in journalism from one of the new journalism programmes in Ethiopia (Skjerdal and Ngugi, 2007) and is convinced that good journalism should be free of interference.He reflects on his work situation in ENA: 'You're a journalist only in verbal terms.You're not critical, you don't ask questions.It's not as such a good experience.People in the private media may picture you as propagandist' (ENA reporter, personal interview, 17 November 2009).

Unspoken Expectations
The editorial policy is frequently mentioned by managers in the news organizations when asked to explain why reporters need to put constraints on themselves in reporting.However, newcomers are not always duly introduced to the contents of the policies.'Nobody has given me the editorial policy,' imparts an editor in ETV (personal interview, 2 December 2008).After a while in the organization he felt a growing need to see the document that was so often referred to by peers in the newsroom, and he asked for a copy.Having studied the policy document he came to the conclusion that it actually didn't put any barriers on his journalistic activity.'The editorial policy doesn't restrict anything.It allows you to do anything,' the editor explains.
The discrepancy between formal editorial policy (of which there is one document for each state media organization, but they are largely similar in content; (Ethiopian News Agency, 2003;Ethiopian Press Agency, 2003;Ethiopian Radio and Television Agency, 2005) and actual journalistic practice is worth a closer scrutiny.The editorial policy makes clear that the foremost aim of the state media is to serve the public.Moreover, recognized journalistic principles such as accuracy, fairness, balance and protection of sources are maintained as nonnegotiable values.Nowhere does it say that one political grouping should be privileged above others, although the state media are assigned a special obligation to inform the public about the government's policies and actions.How, then, should one explain that the reporting is overwhelmingly favourable towards the government and the ruling party?
The answer seems to lie in the unspoken expectations that permeate the news organization.At every level in the organization there are compelling reasons to keep reproducing the loyalty to the owner that has characterized the state media organizations long before the present regime.As such, the interconnectedness between the media and the ruler of the day might to some extent be seen as an inheritance of the media standards that existed during the Derg regime  and the preceding Empire.4 Reporters in this tradition tend to see themselves as civil servants as much as journalists, thus Ralph Akinfeleye's (2005) phrase 'journalism of the civil service rule', which he uses interchangeably with 'protocol journalism', that is, the type of journalism that simply reproduces official information and requires total loyalty to the state bureaucracy.Over the past years, 4 This is not to underestimate the unprecedented propaganda machinery that existed during the Derg regime. Oe senior media practitioner who worked for Radio Revolutionary Voice of Ethiopia (the state radio) in the 1980s elucidates the difference: 'I could not say that I was a journalist in the Derg regime.I was a clerk or something like that.Today, nobody can control me.
[…] No one is coming to my office to ask why I wrote this and that.There is a miraculous difference between the Derg and the EPRDF' (private media publisher, personal interview, 3 December 2008).
however, this tradition has been seriously challenged by demands for journalistic professionalism, in which independent reporting is a core value.Adhering to these new values is also considered an important guiding principle for a reasonable number of journalists in the state media.One editor in The Ethiopian Herald explains how she immediately reacts if she finds that one of her subordinates hands in a sunshine story with blunt promotion of the government: If that kind of article comes to me, I will take proper action.There are some who want to exaggerate what the government does.I don't have tolerance for such material.I have to follow my own standards.(Editor in The Ethiopian Herald, personal interview, 26 November 2008) The concerned editor speaks warmly of internationally recognized journalistic standards and believes the state media, too, should promote the rights of citizens, act as a watchdog and expose failures of the government.In the same interview, however, she admits that 'we're not releasing what we see and hear from the government; we're rather softening it'.This contradiction is again a consequence of the editorial policy, the editor contends, and she goes on to summarize the gist of the editorial policy as such: 'Every state journalist is expected to write from the angle of government interest' (personal interview, 26 November 2008).Once again, this demonstrates how the perception of the contents of the editorial policy and the actual policy document are in conflict.According to Bereket Simon, Ethiopia's minister of government communication, it is the journalists' own problem if they do not understand and adhere to the editorial policy.Contrary to the perception of the journalists, he doesn't see why reporters should be afraid of reporting critically on the government and public administration.'If a journalist tells me he's afraid of that, it's not the problem of the government, but of the individual journalist.Journalists should know their craft' (Bereket Simon, communication minister, personal interview, 10 May 2010).Even so, the journalists interviewed in this study are united in their view that working for the state media means that they have to report positively on the government, and a lesser or larger degree of self-censorship is deemed necessary for this to be achieved.
The paradox, as exemplified by the account of the editor in The Ethiopian Herald quoted above, is that many of the journalists defend professional journalistic standards and claim to be following these standards while at the same time allowing self-censorship.One might perhaps think that a strong conviction of the value of independent reporting and rejection of interference from outside parties would mean that reporters would not accept any degree of self-censorship, but somehow they seem to reconcile these two apparently contradictory philosophies and practices.One possible explanation could be that incidents of self-censorship are irregular and can be seen as a slip in terms of ordinary practice, which could be forgiven as soon as tensions were reduced and political conditions went back to normal, much as Israeli reporters return to independent reporting after a few days of extraordinarily pro-national reporting during a perceived national crisis (Zandberg and Neiger, 2005), However, in the Ethiopian state media, selfcensorship is not provoked by an extraordinary situation but is part of a daily routine.The individual journalist is thus left with the dilemma of reconciling selfcensorship and acceptable journalistic standards without being overwhelmed by professional contradictions and possibly conflict over their personal integrity.
What I shall suggest in the following is therefore that consistently applying selfcensorship involves some degree of justification -or rather semi-justification -of the practice.On the basis of the interviews with journalists in the Ethiopian state media I thus propose four underlying ethical justifications for performing selfcensorship, relating to four dimensions -organizational, audience, professional performance and national ones.These are a further elaboration of a similar set of justifications suggested in an earlier study carried out among Ethiopian journalists (Skjerdal, 2008).The recent restructuring of the state media, however, has intended to specialize and professionalize the different sections of the media, thus perhaps reducing the practice of self-censorship in the more proficient parts of the media.The specialization is seen most notably in the opposing routes that the government decided for ENA and ETV in 2009.On the one hand, it was decided that ENA would be brought directly under the auspices of the newly established Government Communication Affairs Office, meaning that journalists in the organization were more prone to view themselves as information workers for the government; while ETV, on the other hand, merged with Ethiopian Radio and was determined to become more professional in various ways.The latter was reflected for instance in the new policy of making use of the branding 'public media' instead of 'state media'.It is all the more striking, however, that routine habits such as selfcensorship continue to characterize the different institutions and sections despite organizational restructuring and formal professionalization.I thus suggest that the following four justifications are applicable to journalistic reasoning across all three organizations in the study (ETV, The Ethiopian Herald and ENA), regardless of the institutional refinement that has taken place over the past two years until 2010.

Relegation of Ethical Responsibility
The first justification for self-censorship is rooted in the organizational structure and regards the locus of ethical responsibility in the state media.It is suggested that reporters find it less problematic to deviate from professional standards because the ethical responsibility has been removed from the individual journalist to the organization at large.The informants in the study repeatedly indicated that there is a discrepancy between their own journalistic standards and the expectations of the media institution.One way to come to grips with this is to realize that they are not journalists in the ideal sense but rather servants of a state organization, hence justifying the sidestepping of proper journalistic standards.'I don't see myself as a professional journalist, but as a government worker,' admits one reporter (ENA employee, personal interview, 27 November 2008).One of his colleagues concurs and indicates that his own view of ideal journalism is gathered from a wider journalistic community: 'I've been covering international conferences.Only on those occasions do I feel like a journalist, otherwise not' (reporter in ENA, personal interview, 4 December 2008).The realization that the state media has a double set of obligations thus serves, perhaps ironically, as a justification for dubious ethical practices such as self-censorship.
Some informants also suggest that the state media system, by its very structure, overrides the dominant ethical norms of the professional community.It is therefore regarded as useless to strive for more professional standards as long as the government has a firm grip over the media organization.This view is explicated by an ETV editor: 'They [the managers] are not driven by cheap propaganda.But the system, you know, forces you to get involved in that' (personal interview, 17 November 2007).This way, 'the system' becomes an excuse for easing personal ethical standards.

Elasticity of Journalistic Editing
Where does ordinary editorial decision-making end and self-censorship begin?The boundaries are not clear-cut.Leaving out segments of a news story does indeed belong to the very basics of any journalistic activity.The elasticity of journalistic editing could also prove to be a convenient way to justify soft self-censorship.The journalists conferred with in this study are indeed found to be operating within what they consider to be accepted limits of journalistic liberty, at least most of the time.Several informants underline that they would never agree to perform journalistic activity that involved sidestepping fundamental ethical values.One such value is the dispatch of direct falsehood.'We don't lie,' imparts an ETV editor, 'but we hide facts' (personal interview, 17 November 2007).Hiding information is therefore regarded as being within the elasticity of editing, but communicating false information would be considered a professional and personal no-go area.A few of the informants, however, admit that they have been involved in using fabricated or twisted case stories to illustrate a news item.An ENA reporter uses the example of habitually employing the phrase 'dwellers of Addis Ababa agree that …' when in reality the reporters have not been out on the streets actually asking people.Nonetheless, the impression from the field research is that most journalists have a hard time accepting that fabricated or false information is included as part of their news stories.They experience, however, that journalistic editing gives them a lot of freedom to slant the story in favour of the government agenda and still be within the limits of acceptable journalistic practice.
Indeed, the ethical problem associated with biased reporting is not primarily an issue with the single article or story, but with the overall inclination in reporting practices over time.For the individual reporter, the responsibility is mainly to do with the single story, which allows for a considerable degree of journalistic judgement.The elasticity contained herein opens up space for a certain degree of permissible self-censorship.A problematic aspect of this view of journalism, others would argue, is that it leads to a conflict of interests, causing journalists to hide information to avoid harm for the country or citizens.Confronted with this dilemma, a reporter in ENA still maintains that he should at all times be loyal to that which serves development purposes better.His personal assignment in ENA is to cover Ethiopian Airlines, the state-owned air carrier.If he discovers anything that could be harmful to the airline, he will give them a call rather than reporting it in the state media.'In my view, it is better to tell them to take action than to broadcast the news to the public' (personal interview, 4 December 2008).The underlying rationale is that journalists are in possession of a greater knowledge base that sometimes needs to be filtered before it reaches the public -thus opening the way for self-censorship.

Confidence in Critical Audiences
This line of thought is familiar to many other media environments in Africa too.Chief editor Ignatius Kabagame of New Times in Rwanda, for instance, notes that the media in the country 'feel a need for social responsibility through selfcensorship' (Africa's News Media, 2006).Researcher Andrew Kanyegirire (2007), on the other hand, points to the ethical dilemma that has arisen when governments on the continent use the media to portray the New Partnership for Africa's Development (NEPAD) in a positive manner for the sake of the greater good.Journalists he has interviewed are concerned that this approach leads to 'selfcensorship, propaganda, praise-singing and cheer-leading -all of which is in contradiction to the values of libertarian journalism ' (2007, 276).In the Ethiopian state media, however, it appears that the government has received notable support for its development journalism strategy with its own journalists, although not to the extent that they allow the bypassing of standard ethical codes on a permanent basis.That direct self-censorship is not accepted as a regular procedure is illustrated by one editor who mentions the example of incitement to violence in the Rwandan media in 1994 (Thompson, 2007b) as an exceptional case where he could justify self-censorship.'If I had to report such issues, I would report responsibly,' he imparts (ETV editor, personal interview, 17 November 2007).

Concluding Remarks: Discourses of Fear
To sum up, the study has unearthed widespread self-censorship practices in the Ethiopian state media.The adoption of the practices takes place when new reporters are introduced to the organization and learn the 'official' reporting style as well as which stories are suitable for coverage in the state media.Self-censorship is, for the most part, viewed as an unfortunate practice by the journalists, and they apply various justifications to validate the practice.Four such justifications have been identified in the study: (1) relegation of ethical responsibility; (2) elasticity of journalistic editing; (3) confidence in critical audiences; and (4) adherence to social responsibility.The study has also pointed to a significant degree of discrepancy between the relatively open-minded editorial policy of the state media and the restrictive reporting practices.
That self-censorship practices persist despite developments towards professionalization in the state media organizations may be attributed to a feeling of uncertainty and discourses of fear among the journalists.When asked about the reasons for restricting themselves in reporting, the journalists typically refer to the fear of losing their job or getting other types of punishment.However, interestingly enough, only a few informants are able to name specific journalists who have received any specific punishment for transgressing editorial or ethical codes.A journalist for Addis Zemen, a state-owned magazine, characteristically imparts that 'there are reactions in the form of fines and salary cuts, I've heard from senior journalists' (personal interview, 4 May 2009); but, when trying to think of specific cases, he is not able to name any.
Uncertainty is a fertile ground for self-censorship, and is perhaps more significant than media laws when it comes to imposing restrictions on journalists.Greg Simons and Dmitry Strovsky (2006) demonstrate that, in Russia, censorship has shifted from a hierarchical and predictable form during communism to become much more personal today, and that self-censorship is heavier than it used to be (cf.Kenny and Gross, 2008).In Ethiopia, the exact degree of self-censorship is difficult to measure, but the degree of uncertainty among journalists in the state media is significant.It follows from the argument in this study that future research on censorship practices in semi-democratic societies like Ethiopia is likely to gain more from in-depth studies of newsroom behaviour and journalist perceptions than merely the documentation of official censorship practices and a scrutiny of censorship laws.
What you have in the state media is spoon-fed journalism.When you come in the morning, there are quite a lot of press releases on your desk.You distribute them and send out the reporters.It's easy work because you know what the day is going to look like.(Editor in The Ethiopian Herald, personal interview, 4 May 2009) (Government of Ethiopia, April 2008) a product of internal in-house judgements, but of audience relations as well.In Ethiopia, probably typical of state-controlled media societies, journalists sense that the audience is able to see through the rhetoric of the state media.'Wearebelievedbymany to be mouthpieces of the government,' says one ENA editor characteristically(personal interview, 21  November 2008).This makes it easier for journalists to justify unethical practices like biased reporting and self-censorship based on the argument that professionalism is not expected to be a feature of the state media anyway.Many journalists have a pessimistic view of audience confidence in the state media, as illustrated by the following confession by one of the editors in The Ethiopian Herald: 'Sometimes I'm exhausted and I leave the story without editing.But nobody reads it anyway' (personal interview, 24 November 2008).Critical audience perceptions of the Ethiopian state media are confirmed by several case studies, such as BadesoHaji's (2008)reception study of ETV programmes among Afan Oromos, which found a generally low audience satisfaction rate, and Daniel Bekele's (2008) study of the same station, which found that it fails to deal satisfactorily with audience responses and complaints.If the situation was reversed and the audience had high confidence in the state media, journalists would have found it more difficult to justify self-censorship because they would feel they were deceiving the public.Adherence to Social ResponsibilityThe fourth justification of self-censorship identified in Ethiopian newsrooms is related to the journalists' normative view of the media.They generally support the idea that the media -especially the state-owned media -should play a pro-active role in nation-building.To this end, self-censorship is sometimes justified as a necessary strategy for performing socially responsible journalism.It is argued that journalists should see themselves as social change actors on par with other agents in national development.This view of journalism fits in with the official editorial line of Ethiopia's state media, which foregrounds a development journalism philosophy.The philosophy is laid out in an extensive policy document dated April2008(not yet formally ratified), which describes development journalism as 'a journalism that makes people understand, accept and actively participate in the implementation of appropriate developmental ideas that may extricate people from poverty and backwardness by bringing about rapid national change'(Government of Ethiopia, April 2008).These ideas have a certain resonance with the media workers as well.'Journalism should be an instrument of change,' says one Ethiopian Herald editor (personal interview, 24 November 2008).Another editor (of ETV) sees no problems with using current affairs programmes to attract investors to come to Ethiopia.He gives the example of a documentary that he produced about the Ethiopian horticulture industry, which was transmitted on satellite to foreign audiences.'It is nice for the government and nice for the country.It is nice that foreigners can come and invest,' says the editor (personal interview, 2 December 2008).