Distributing ethics: Filtering images of death at three news photo desks

This article explores the practices of selecting news images that depict death at a global picture agency, national picture agency and a news magazine. The study is based on ethnographic observations and interviews (N = 30) from three Western-based news organisations, each representing a link in the complex international news-image circulation process. Further, the organisations form an example of a chain of filters through which most of the news images produced for the global market have to pass before publication. These filters are scrutinised by the empirical case studies that examine the professionals’ ethical reasoning regarding images of violence and death. This research contributes to an understanding of the differences and similarities between media organisations as filters and sheds light on their role in shaping visual coverage. This study concludes that photojournalism professionals’ ethical decision-making is discursively constructed around three frames: (1) shared ethics, (2) relative ethics and (3) distributed ethics. All the organisations share certain similar conceptions of journalism ethics at the level of ideals. On the level of workplace practices and routines, a mixture of practical preconditions, journalism’s self-regulation, business logic and national legislation lead to differences in the image selection practices. It is argued that the ethical decision-making is distributed between – and sometimes even outsourced to – colleagues working in different parts of the filtering chain. Finally, this study suggests that dead or suffering bodies are often invisible in the images of the studied media organisations.


Introduction
Natural catastrophes, wars and conflicts are events of human suffering that are regularly covered by the media. The way these events are represented to citizens is largely dependent on media organisations, who act as gatekeepers for visual coverage. While traditional gatekeeping remains a valid practice, it may have lost some of its relevance and changed more to gatechecking (Schwalbe et al., 2015) and providing explanations and context (Standaert et al., 2019).
For example, the 'Black Lives Matter' movement has been spreading its message against racial discrimination and violence again since May 2020, after the death of George Floyd. In this case, the social media images and videos posted by ordinary citizens, activists and others helped elevate the topic to become a part of the global news agenda and mobilise people to protest. This resulted in the bypassing of the traditional gatekeeping processes of professional media and made the incident visible everywhere regardless of racial politics. Thus, as professional media no longer have full control of the publicity of such events, they constantly reposition themselves in relation to the social media representations (Blaagaard et al., 2017). In contemporary reporting, the citizens' voice appears alongside the journalists' and is treated as a source of information; not only as an eyewitness's testimony or emotions (Chouliaraki, 2015a: 106). Furthermore, according to McKinley and Fahmy (2011), the public can handle more shocking images than before because of their increased exposure to graphic material through online forums. This development has created emergent ethical dilemmas for the gatekeeping processes in media organisations. Media professionals cannot ignore social media visibility, but they need to act according to the journalistic values, such as objectivity and truth-telling, that are particularly crucial when it comes to graphic images (see also Nilsson, 2020: 271).
The extent to which graphic images should be present in daily news media has been a source of continuous debate both in journalism research and in practical work (Hanusch, 2013;Morse, 2014). Earlier research has found that the discussion surrounding these images is based on two fundamental values of journalism: providing information and respecting the victims (Morse, 2014). Both principles potentially have equal moral legitimacy in journalism (Silva and Eldridge, 2020: 31-33), and, therefore, professionals need to justify their decision in each situation.
In this paper, I present an empirical example based on interviews and newsroom observations of photo editors and other photojournalism professionals who handle, evaluate and disseminate news images that are eventually published in news media. The work of these professionals often remains hidden, even though they are powerful actors in the journalistic decision-making process when it comes to the practices of image selection and visual coverage. Gürsel (2017: 46-47) calls these professionals 'image brokers' and suggests that they not only determine the visual coverage that we see but also participate in building our visual realities. This is done in the everyday practices, for example, by arranging and sequencing pictures into balanced binaries that eventually reproduce normative worldviews.
This study explores the ethical reasoning of photojournalism professionals from three media organisations with respect to images of violent death. Each of these organisations is a part of the chain through which news images travel from the global market to a national market and, finally, to an end-user making publication decisions. The data for this study was gathered from (1) a global picture agency Reuters Pictures (hereinafter RP or Reuters), (2) the Finnish picture agency STT-Lehtikuva (hereinafter LK) and (3) the Finnish news magazine Suomen Kuvalehti (hereinafter SK).
A central idea of this study is the notion that the exact same images often travel through all three levels -global, national and local end-user -and each organisation adds different layers of filtering. Thus, some images are filtered out at each stage. While the photo editors of the three organisations often work with the same images, the business environment and customer expectations are different in each case. The common aspect for all three is that they are all Western-based commercial news organisations. They are based in the United Kingdom and Finland, where there are relatively high levels of press freedom, professionalism and economic development, which makes them similar enough to study the filtering chain. Given the ethnographic nature of this research, the three organisations are examined in parallel instead of being strictly compared.
Earlier research has recognised professional journalism's ambivalence between the professionals' commitment to social responsibility as a form of public service and the pressures faced by the business-oriented organisations (Christians et al., 2009: 54). In this study, the ideological ethics of journalism will be explored against the different organisational and political agendas in a Western context. Given the aforementioned ambivalence, this study seeks to determine whether there are many different ethical norms in the field of journalism. The research addresses three fundamental questions: RQ1: How are ethical norms constructed in the discourse of photojournalism professionals? RQ2: What are the similarities or differences in the photojournalism professionals' ethical reasoning regarding images of death at a global picture agency, a national picture agency and a local news magazine? RQ3: How do the different layers of filtering (global, national, local end-user) affect the published visual media coverage?

Gatekeepers' ethical orientations
The term 'gatekeeper' was coined by Lewin (1947) in the context of group dynamics and was first introduced in journalism research by White (1950). In newsrooms, gatekeepers have traditionally controlled the news flow by deciding what to publish and what to leave out (White, 1950). As for selecting news images that is, visual gatekeeping, there are many steps, as the images are evaluated by different professionals in the production chain. Usually, the first 'gates' are the senior editors, who decide what will be photographed in the first place. Then, decisions are made by the photographers, who send a selection of images to the photo desk, where the final selection is made by the photo editors (cf. Schwalbe et al., 2015: 466). Today, gatekeepers also actively search for and verify social media content (Tandoc and Vos, 2016).
In their study, de Smaele et al. (2017) investigated visual professionals' image selection practices at a newspaper's photo desk in Belgium to find that the most important criteria for selecting an image was its capacity to convey the news. They highlighted that, beyond the individual level, image selection is highly affected by organisational and practical constraints and the overall professional environment. Similarly, a study focusing on American newspaper photo editors' decision-making after the 9/11 attacks found that the editors emphasised the images' ability to communicate the story while being concerned about the privacy of the victims and the readers' response (Kratzer and Kratzer, 2003). Furthermore, Fahmy (2005) surveyed American photojournalists' and photo editors' decision-making regarding 9/11 and the Afghan War to find that the decisions varied depending on context and that political factors potentially shape ethical decisions. In a similar vein, Nilsson (2020) suggested that proximity and cultural factors were of significance when Swedish newspapers decided to withhold graphic images after a hijacked truck crashed into crowds, killing and injuring people in Stockholm in 2017. According to Nilsson, the newspapers aimed to appear ethical in contrast to those on social media.
When studying the influences that affect the ethical orientations of journalists, Hanitzsch et al. (2010) found that different theoretical models and empirical findings often contain similar levels of influence, such as that of individual/role, organisational, medium/industry/institution and cultural/ideological factors. However, these are conceptually overlapping, and different models place the influences on different levels. Using a multilevel analysis that considered the different influences in a cross-national setting, Plaisance et al. (2012) approached journalists' ethics from two moral philosophical orientations: idealism and relativistic thinking. They found that professionalisation and the existence of codified editorial rules seemed to be determining factors for the adherence to idealistic thinking among journalists. Whereas, it was more difficult to identify the factors that explain the relativistic thinking of journalists (Plaisance et al., 2012: 656).
This study reflects on the different influences and provides empirical examples of the relations between the professional ideology and relativistic thinking of photojournalism professionals. A qualitative approach to the micro and meso levels of photojournalistic work was chosen to explore the ethical decision-making process on a more practical level since in survey studies, the ethics often remain more at the level of ideals (Örnebring et al., 2016).

Visual news coverage of death
Negative news, such as of conflicts or other events where people die or are injured, often have a high news value. The news value of such events is related to their societal significance, which increases with the potential impact of the event and the number of people involved (Harcup and O'Neill, 2017). Thus, it is crucial for the media to cover such events both textually and visually. With death scenes and other graphic news events, news images serve as eyewitness reports and illustrate the photographer's on-site presence (Zelizer, 2007). Despite the fact that photographic meaning is easy to manipulate, news images are treated as evidence in journalism (e.g. Hanusch, 2010: 57).
For photographers and photo editors, there are numerous ways to cover such events; the chosen strategy depends on many cultural and political aspects as well as the motives behind the publication. Earlier research has found that there are differences in the coverage of death between countries or media systems (see Hallin and Mancini, 2004) and types of publications (Hanusch, 2013). For example, when comparing newspapers in Western Europe and North America, Hanusch (2013) found that tabloids published more death images and that those images were often more gruesome than those published in the broadsheets. Furthermore, the differences between tabloids and broadsheets are more pronounced in democratic corporatist countries than in liberal countries. The broadsheets in Northern European countries tend to censor graphic images more than those in the United Kingdom, United States and Canada. In contrast, the tabloids from Northern European countries publish more gruesome images than their liberal counterparts (Hanusch, 2013).
Media scholars have also heeded to the notion that the geographical and cultural distance of an event affects the level of graphicness of the published images (Hanusch, 2010: 41-45;Morse, 2014). If the news image depicts proximate instead of distant suffering, a sense of national loyalty takes over the journalistic practices (Nossek, 2004). The proximity of a violent event has resulted in, for example, professionals self-censoring the images by using long shots instead of close-ups and subtler images instead of explicit ones (de Smaele et al., 2017: 64;Morse, 2014: 107-109).
Furthermore, according to Hanusch, different studies from Western countries have found that death and especially graphic death are rarely seen in the Western news media. This means that even though deadly events are often reported by them, graphic elements, such as blood, corpses or body parts, are rarely depicted in the news images (Hanusch, 2010: 59-63). War, in particular, is a difficult topic for news media because it is subject to contradicting levels of interest. Scholars have demonstrated, for example, that since the 1990-1991 Gulf War, the American news media has been reluctant to show injuries or deaths from the war zones where American troops are involved (Griffin, 2004).

Material and method
This study is a part of a larger research project, and the material has been gathered by interviewing photo editors and other photojournalism professionals from the three newsrooms mentioned previously. The interviews were augmented by newsroom observations in the Finnish newsrooms. In the case of Reuters, the choice of interviews, instead of observations, was a matter of access.
The picture agency Reuters is one of the leading news picture agencies in the world. At the time of gathering the data, Reuters had two global offices -in Poland and Singapore -and most of the photo editors worked out of these offices. Reuters' global pictures editor is based in London. The material for this research was gathered at Reuters' London bureau and at their global office in Poland.
Similarly, the national agency LK is the largest news photo agency in Finland; it cooperates at some level with almost all the Finnish daily news media organisations. LK is one of the customers of Reuters. The agency's primary clients are Finnish professional media organisations; in addition, LK exchanges images with international picture agencies. When reviewing foreign news coverage, the photo editors' job is to find an angle that has national interest. Further, when providing pictures for the international market, the Finnish photo editors consider whether the local event would have international appeal.
Finally, the news magazine SK was among the first magazines to begin publishing photographs in Finland. Unlike the picture agencies, the magazine is a publisher instead of a wholesaler. SK has access to the image feeds of many different picture agencies, including the Finnish agency LK, which provides images produced by global agencies such as the Reuters. SK's readers have been characterised as being middle-aged or older, educated and a part of the elite. The magazine is mainly sold by subscription. The SK covers world news events, international politics and related phenomena in Finland.
The data consist of a total of 30 semi-structured interviews with photo editors and other photojournalism professionals from all three organisations in addition to field notes from participant observations within the newsrooms of LK and SK (see Table 1 for a detailed overview of the data gathering). During the fieldwork, I attended the photo department's daily morning meetings at LK and weekly meetings at SK and usually spent the rest of the day observing one or two photo editors at a time. During the observation, I posed questions or requested them to explain what they were working on. For example, a typical question was asking why they decided to select a certain image.
The interviews at LK and SK were conducted during and immediately after the observation periods in the spring of 2017. The interviews at Reuters were conducted in the spring of 2018 in London and in the autumn of 2018 in Poland. The total participants comprised 19 men and 11 women. Their work experience ranged from less than 5 years to more than 30 years in the field. At LK and SK, the respondents were all Finnish nationals, while, at Reuters, they represented many different, but mainly, European nationalities.
The interviews lasted between 33 and 90 minutes each and were recorded and transcribed. The interview respondents were consecutively numbered according to their organisations (RP1, LK1, SK1, etc.). The pre-defined themes of the interviews, based on my prior empirical work in newsrooms, were as follows: daily workflows and routines, rewarding and challenging aspects of the work, image selection practices and criteria and professional ethics. This article focuses on the data regarding image selection and the professionals' reasoning behind using images of death. The topics were approached with questions such as, 'What criteria do you use when selecting the news images?', 'What ethical challenges do you face in your work?' and 'Where would you draw the line on graphic images?'. After gathering the data, the material was qualitatively analysed. I coded the interviews using the Atlas.ti software and conducted close readings of the interviews in parallel with the field notes. The interview excerpts and field notes, wherein the participants described the ethical challenges they face in their work and talked about images of violence or death, were selected for the analysis. The thematic analysis was used for organising the data into thematic clusters (Nowell et al., 2017). In the first phase, two major thematic clusters were identified: (1) arguments for informing people and (2) arguments for restricting certain images. Professionals at each organisation had somewhat similar opinions about these two broad themes at the level of ideals and principles. In addition, there were smaller thematic clusters for which the material was more diverse. These include (1) preferences of clients or audiences, (2) national legislation and cultural conventions, (3) geographical or cultural distance and (4) availability and quality of the images.
Through the final analysis, I created three thematic frameworks through which photojournalism professionals discursively construct their professional ethics: (1) Shared ethics: People's right to know versus respecting the victims (2) Relativistic ethics: Ideals versus practice (3) Distributed ethics: Teamwork versus outsourcing

Findings
The three thematic frames that inform photojournalism professionals' ethical decisionmaking will be introduced subsequently. The examples reveal the global, national and end-user levels of filtering, which will be further explored in the conclusion.

Shared ethics: People's right to know versus respecting the victims
In line with earlier research (Morse, 2014), the professionals from each organisation referred to the importance of ideological ethical values, such as journalists' responsibility to provide information to people about what is going on in the world in an ethical manner. These values have been acknowledged in journalism's self-regulation and have been stated in many ethical guidelines of Western journalism. For example, the American Society of Professional Journalists (SPJ) has compiled the central journalistic values in their Code of Ethics. 1 Minimising harm is one of the cornerstones of SPJ's ethical practice and the guidelines formulate that 'journalists should balance the public's need for information against potential harm or discomfort'.
The interviewed professionals assessed whether certain images should be published or circulated by evaluating the impact and significance of the photographed event in relation to the potential harm that its publication might cause. Further, the level of graphicness has to be in line with the significance of the real-world event.
If the event is very gruesome, such as war, then we will provide gruesome images. (LK9) Behind this reasoning lies the demand for coherence in relation to the real-world event as well as the relation to the other published imagery and visualisations that news media has provided for similar previous events. Meanwhile, the professionals considered it important that the reporting be accurate and ethical. That is, while providing information, it is also necessary to filter images based on ethical as well as practical grounds. We try and be respectful of the people in the pictures, whether they're victims or collateral people, relatives, whatever. But at the same time, the news has to get out there. We don't wanna be censoring ourselves. (RP18) These quotes highlight both aspects of ethical reasoning that the photojournalism professionals must strike a balance between. Moreover, these quotes reveal that the participants are concerned about the depicted subjects, their next of kin and the wider news audiences. In comparison, earlier research has found that the photojournalists who work with photographic subjects care more about the victims than the news audiences (Miller and Dahmen, 2020).

Relativistic ethics: Ideals versus practice
Many participants emphasised the need for case-specific consideration when selecting or rejecting images. They often referred to practical constraints, such as the idea of balance, availability and quality of the images, client or reader expectations and cultural conventions, as notable factors in their decisions. The reasoning informing these practical questions is more relativistic than based on journalistic ideals.
For example, as a global picture agency, Reuters aims to provide a large variety of images that illustrate different perspectives. According to the interviews, the main ethical concern for Reuters is of not coming across as biased in its coverage. This is also emphasised in their ethical guidelines. 2 As a practical solution, Reuters sends photographers to capture both sides of a prolonged conflict. As a result, the image selection process for such issues is influenced by balancing between the casualties of the opposing sides. Providing neutral reporting is also a part of the agency's captioning and keywording practices. Biased words, such as 'terrorist', will be replaced by more neutral terms, such as 'fighter', in image captions. Earlier research has also acknowledged a similar tendency of avoiding taking a stand in editorial language (Barkho, 2019). The maintenance of neutrality and balance ensures the news agency access to conflict zones. Thus, the choice of whose suffering will be represented in news media is not solely determined by the ideological values of journalism nor the qualities of the images but the very structures that make the production of those images possible in the first place. Meanwhile, the idea of balance often comes to stand for journalistic objectivity (Gürsel, 2017: 41-46). Furthermore, the participants stated that much depends on the available images and the whole that these images form. One participant described their choice of a series of images from a plane crash in Nepal (Figure 1), where they decided not to use images of dead bodies even though such pictures were available: We have the wreckage of the plane, we have some wider view of this horrible disaster, but we don't need the picture of a burned body in close-up because that doesn't give anything apart from the shock value, which we don't want to stand for. (RP20) The influence of availability on image selection practices has been recognised in earlier research too (Nilsson, 2017). The professionals from all the organisations also used the term 'shock value' as something that they wanted to avoid (LK1, SK11, SK12, RP20). This term usually refers to images that are considered distasteful, intrusive or just morally wrong. The shock value of images was associated with sensationalism that lacks societal importance. Furthermore, in cases such as the plane crash, one can easily imagine the repercussions for the passengers in such a disaster, and, therefore, a picture of a burned body would not add anything but a shock value to the reporting.
Moreover, respecting the victims and their next of kin was considered especially crucial in local coverage where it may be possible to identify the victims. One photo editor mentioned the example of a local bus accident where many young people were killed in 2004 in Central Finland (Figure 2): The first images that we received were full of dead bodies. Recognising one's own child was possible, and this was before the parents had been informed about the tragedy. I could by no means pass on these pictures, especially as it [ This quote underlines the focus on and special consideration of local news events and describes how cultural or geographical proximity affects the photo editors' decisionmaking. As earlier research has demonstrated, deaths in different parts of the world and among different groups of people are not represented equally by news media (Hanusch, 2010: 37-45). According to Chouliaraki (2015bChouliaraki ( : 1368, there are 'historical asymmetries between the dignified invisibility of dead of the West and the indignity of its visible "others"'. Feldman (1994) called this 'cultural anaesthesia' referring to the media representations of the other. This could be interpreted here as media's blindness towards its own conventionalism ensuring that media images do not deviate excessively from the 'normal', where distant suffering is represented more explicitly than proximate suffering. Campbell (2004: 64) argues that the danger of othering lies in the tendency to treat domestic suffering as superior in comparison to the foreign, thereby strengthening racial and cultural stereotypes. Furthermore, he stresses the importance of context and argues that media themselves regulate the 'taste and decency' of representations of death. The practices of display are important, since images are part of media narratives together with textual arguments and sites of presentation (Campbell, 2004: 70).
A few Finnish participants also mentioned that usually they do not mediate the most graphic images from international picture agencies to the Finnish clients (LK2, LK6). This was explained by client expectations and cultural conventions (see Nilsson, 2020: 270 about Sweden). One Finnish photo editor recalled a conversation with a European newspaper client about images from the sinking of the MS Estonia cruise ship in 1994. The ship was traveling from Tallinn, Estonia to Stockholm, Sweden when it sank off the coast of Finland, killing more than 800 people: These two Finnish examples reveal that LK's clients constitute a more homogenous whole than those of Reuters. The photo editors seemed to share an understanding about the kind of images that would likely be published by the Finnish news media, and this influences their image selection process. In addition, one of the participants stated they do not provide the most graphic images to protect the audiences, especially young children. 'I think we mainly try to protect children . . . so that a child would not see such an image' (LK6) (also Keith et al., 2006). This line of thinking, however, is not imperative. The professionals also provided examples of cases where images of death were used. These included images of a murdered politician, images of gas attacks from a war zone and images of death from natural catastrophes from different parts of the world. For example, both the studied picture agencies circulated the image of the drowned Syrian refugee child, Alan Kurdi, lying face down on a Turkish beach in 2015. One Finnish editor commented on their decision: We knew that it was an image that we had to get to our clients almost at all costs, because it was clear that it was going to be a 'big ' image. (LK1) This quote reveals the pressure on the editor to intermediate the image, since it was becoming viral on social media. In this case, the decision was weighted against to the fact that the image was already 'out there', and the professionals felt that they had no choice but to circulate it. Despite the horrifying fact the image depicted, it allowed identification and has been interpreted to fit to the Christian visual tradition, especially into that of a sleeping angel (Drainville, 2015) (for problems of beautification, see Reinhardt, 2007: 28-31). I argue that the way death was represented in the image, furthered the professionals' decision in this particular case.
Furthermore, the Finnish picture agency provided numerous death images from the South Asian earthquake and tsunami of 2004, which was one of the deadliest catastrophes in recorded history. Many Finns also died or were otherwise involved in the catastrophe. The photo editors justified their decision for providing such images by referring to the magnitude of the disaster and their duty as information providers. However, it was, again, taboo to show images of Finnish victims in the national media.
Moreover, a few editors mentioned that, sometimes, the reason for providing images of death is to motivate people to take action (LK3, LK5, LK6). One senior photo editor recalled the Civil War in Rwanda in the beginning of the 1990s; he had provided graphic images of the conflict because he felt that it was something that people needed to know about.

I had thought that if I pass on some graphic images, it could motivate some people to give money to the Red Cross, for example, and it could then help the situation. (LK5)
The photo editor's aim to motivate people to take action could be interpreted as a noble mission that goes beyond simply informing people. Furthermore, this indicates that the photo editor had a margin for individual consideration and he could have chosen to restrict the gruesome images.
Furthermore, the image selection process depends on how the scene of death has been photographed. As one participant noted, 'Showing what has happened can be done in many different ways, and some ways are acceptable, and others are not' (SK15). For example, a participant from SK talked about the discovery of a mass grave that the magazine's photographer had photographed during the Syrian Civil War in 2013. A close-up image of a dead man's face partly peeping out from the red soil was published on the cover of the magazine. The reason for publishing the image, according to the participant, was that the 'case became such big news' (SK12). The participant remembered that they had considered publishing the image upside down so that it would not be too shocking for readers, but the magazine ultimately chose to publish it the right way up.
Balancing between realism and art is typical for a news magazine that, on one hand, wants to relate with the documentary news tradition but, on the other hand, allows itself more freedom in its visual style in comparison to newspapers. This example, together with the interviews and observations, revealed that SK's image choices depend as much on the style of the images as on the image content as such.

Distributed ethics: Teamwork versus outsourcing
During the study, it became clear that photojournalism professionals' ethical decisionmaking is about teamwork, where the decisions are distributed among the professionals. In some cases, the distribution of ethical reasoning goes even further, and the decisionmaking is outsourced to the clients.
In the structure of global agencies, the pictures travel from local bureaus supervised by regional bureaus to the global offices, which finally send them to clients who then make publication decisions regarding the images in various local contexts (Ilan, 2019b: 40-148). At Reuters, the participants emphasised the cultural and legislative aspects of different countries, and the photo editors stressed their trust in the local photographers.
Certain countries have certain rules that apply. Let's say, in France, you can't publish a dead body, or it needs to be covered, so it's forbidden. . . But the idea here is that . . . we cannot know the legislative things of every country regarding these topics. . . our people there, that's why they're there, they know these things, so they should know when they can do something or not. (RP21) The above citation reveals the global photo editors' need to balance between the global and local levels of the work (also Ilan, 2019a). Meanwhile, it demonstrates the extent to which the local photographers' knowledge is widely appreciated by the editors at the global office. Thus, the photo editors partly rely on the ethical decision-making of the photographers.
Finally, ethical decisions are gauged against the values and expectations of the clients or audiences. Many participants at each of the researched organisations mentioned this as a guiding principle for selecting images.
So, we're offering this stream of images, knowing that we have different types of clients. Some that will use this picture. We know that many others will not. But we know that this is a valid picture for a lot of our news or some of our news media. We put it out there. If we knew that it was a picture that nobody would use, we wouldn't put it out. (RP18) This kind of business-oriented explanation was also common with regard to images of death. To select the right kinds of images, the photo editors need detailed knowledge about the different clients or readers. A common supposition among the professionals is that they know what their clients or audiences appreciate.
. . .we don't show anything just to shock. That is not part of our policy and our voice. It is in a way a certain respect for the readers in the sense that they don't have to look at it, or they may find another magazine for looking at it [disturbing material]. (SK12) In this quote, the participant clearly distinguishes between their magazines content, which they consider responsible, and any kind of material that may be found, for instance, in other kinds of publications or on social media platforms. Meanwhile, they assume that their readers appreciate the magazine's choices. The interviews and observations revealed that the professionals' alleged knowledge about their clients or audiences does not primarily come from reception research but from their experiences with and perceptions about the behaviour of those groups. It is actually a part of the photojournalism professionals' expertise to monitor the behaviour of the clients or audiences and learn from it (Mäenpää, 2020: 9).
At the scale of global picture agencies, not all images are targeted for all clients because of different local interests as well as cultural and legislative issues. The photo editors are able to tag the photos with different restrictions if they, for example, consider the images culturally sensitive or if the image cannot be distributed to certain countries or clients for business or economic reasons (Ilan, 2019b: 84-90). One participant mentioned an example of an image that would probably divide opinions between the clients (Figure 3) The above example is of a typical situation where the photo editors use case-specific reasoning based on their work experience and their knowledge of the clients. A popular opinion among the professionals was that their case-specific decision-making is not selfcensoring but rather an instance of ethical reasoning, which is an integral part of responsible journalism. Such case-specific considerations allow the professionals latitude to make moral judgments that may vary according to cultural conventions or distance from the event for example.
Moreover, as in the quote, the interviewed picture agency professionals commonly emphasised their role as wholesalers. Except for their online news channels, the picture agencies in question do not make publication decisions regarding the images. When providing graphic images, they add disclaimers to the metadata, noting that the image contains graphic content. Thus, the ethical reasoning regarding whether to publish the image is emphatically left up to the client. One participant stated that the main reason for the disclaimers is that they allow those clients who auto-publish the image feeds of the picture agencies to automatically filter the graphic images. Furthermore, the practice is associated with picture agencies' role as intermediaries. The reasoning presented in the above quote was a common sentiment expressed by the picture agency participants. This frees them from certain liabilities and allows them to provide pictures more freely. Wholesaling practices are evident at Reuters and LK. LK often receives images of death from their international partner picture agencies with disclaimers that the images contain graphic content. The LK then sends the selected images to the local clients with the disclaimers left in place. Thus, this practice of outsourcing the ethical decisions allows the picture agencies to maintain that they are fulfilling their duty as legitimate information providers and that they fully support the people's right to knowledge without self-censoring images.

Discussion
This study explored the ethical norms manifested in the photojournalism professionals' ethical reasoning regarding images of death at a global picture agency, national picture agency and a local news magazine. The most commonly shared logic in ethical reasoning was found at the level of the ideological values of providing information and respecting the human dignity of the victims. This finding appears to reaffirm the similar conclusion of Morse (2014). Meanwhile, the findings suggest that the ideological level is often bypassed by the deliberative daily practices and constraints. In line with earlier research, this study found that the relativistic ethics discourse emerged when the professionals emphasised the work's practical preconditions, such as availability and quality of images, client or reader expectations and cultural conventions, as reasons for selecting or rejecting images (cf. de Smaele et al., 2017;Fahmy, 2005;Kratzer and Kratzer, 2003;Nilsson, 2020).
The main contribution of this study is the notion of distributed ethics and its effects on the different layers of filtering (global, national, end-user). The professionals from all the organisations work collectively and distribute the ethical decision-making among their colleagues in the image production chain. All professionals have their own tasks and expertise; however, it was found that they strongly rely on each other and trust that each professional works ethically. Furthermore, picture agency professionals, who mainly intermediate but do not publish images, outsource a part of the ethical decision-making to their media clients by adding disclaimers and restrictions to the metadata of the images. This allows the picture agencies freedom from certain liabilities and an opportunity to provide pictures more freely.
Distributed ethics has an effect on the layers of filtering and thus, on the published visual news media coverage. The variety of images is greatest at the global level, where the filtering is related to the decisions about what will be visually covered in the first place. Here, the picture agency's local and regional bureaus have a significant influence. The national picture agency selects far fewer images, and, for example, the Finnish LK filters out the most gruesome images provided by the global agencies. However, the use of disclaimers allows for less filtering to be done at both global and national levels. At the end-user level, it is crucial that the images fit the style of the publication and are an inseparably part of its stories. Thus, only a few picture agency images end up being published in the magazine. These layers of filtering support the finding that, at all stages, the image market is determined by the needs of the local markets (see Ilan, 2019b: 90).
Furthermore, this study suggests that dead or suffering bodies are often absent or aestheticised in the images of the studied media organisations (cf. Hanusch, 2010). This may be done, for example, by using long shots instead of close-ups, showing only references of a certain tragedy or by picturing the victims in a way that the scene does not rouse excessive aversion (cf. Reinhardt, 2007). More broadly, journalistic conventions have guaranteed that media representations do not deviate too much from the norm, such as mainstream politics, and, thus, the published imagery rarely challenges existing worldviews (Andén-Papadopoulos, 2008: 8-11). This might be a result of respecting the victims, but, arguably, it also reflects the media organisations' intention to protect and serve their clients and audiences. According to Ilan (2019b: 41), the preferences of clients may sometimes directly influence the future processes of production. Therefore, it is possible that the market logic prevents the publication of the most graphic images.
If death is made invisible or aestheticised in media imagery, the media fails to show the world as it is. This may shape the societal understanding of violent death in a way that people do not feel an urgent need for action, for instance, in cases of injustice. As Reinhardt (2007: 21) writes, referring to Sontag, 'aestheticization prompts passivity or contentment in the face of trauma and injustice'. For example, the widely published Kurdi image did arouse political responses immediately after being published; but, in the longer term, a lasting change in public opinion towards refugees was not evidenced (Burns, 2015). Of course, straightforward implications cannot be expected from the publication of single images. However, as earlier research has suggested, repeatedly images contribute to building our worldviews and visual realities (Gürsel, 2017: 46, 47;Reinhardt, 2007: 13, 14).
Future research is needed as the data for this study was gathered at three Western news organisations in a limited duration, which makes it challenging to generalise the results beyond these particular newsrooms. Furthermore, an important aspect of nonprofessionals infiltrating the professionals' visual gatekeeping is beyond the scope of this article. This calls for future research to further our understanding of the ethical decisionmaking practices regarding the sourcing, verification, authorship and informational quality of social media images and how these are assessed in comparison to professionally produced images of violent death. Meanwhile, the findings of the study demonstrate that future research is required to critically examine the role of photojournalism professionals-from a geographically and organisationally diverse perspective-in order to more comprehensively understand the societal and political implications of the invisibility or aestheticisation of death in news media images.

Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is funded by post-doctoral grants from the Kone Foundation and the Academy of Finland 325160.