The Discourse of Migration in English- language Online Newspapers: An Analysis of Images

This paper examines English-language online newspapers from the UK, the USA and Australia and analyses the images of news articles about migration. To do so, the newspapers The Guardian, USA Today and The Sydney Morning Herald have been chosen and news articles collected in August 2016. The corpus consists of 650 news articles comprising about half a million of words. From these, 1,300 images have been extracted, sorted into migrationand not migration-related images and then grouped thematically. Selected and high-frequency categories are examined in more detail, example images analysed more closely, and patterns found and common characteristics are discussed. Investigating a larger image corpus is rare in discourse analysis and is still missing for migration-related issues.


Introduction
Discussions about migration have been prevalent in the news especially since the last few years. However, the concrete directions of such debates vary from country to country. Across English-speaking countries, for example, the immigration debate of Mexicans is leading in the USA, the so-called "European migration crisis" of 2015 is an important issue in the United Kingdom and the arrivals of people by boats are taking a main role in such discussions in Australia. And yet, all these debates cover migration in a largely restrictive manner, ignoring that migration is not only about refugees and asylum seekers, but includes migrating workers, students or married partners, for instance.
Which images are chosen to accompany, introduce, illustrate articles is also highly interesting. Images can redefine the topic of an article or narrow it down and regard issues only from a certain angle. Past research (e.g. Bateman 2014: 244; Machin & Polzer 2015: 22) has also shown that images play an important role in drawing the attention of readers to the article. The extent of such an effect, however, is controversial. For sure, images are cues that are highly significant in our world that is increasingly becoming more and more visual.
This paper therefore examines images in online newspaper articles about migration of selected Englishspeaking countries. To do so, (critical) discourse analytic methods are used, following Machin and Mayr (2012: 54-55) and especially the analyses of newspaper articles and images carried out by Bednarek and Caple (2012: 163-175), Caple (2013: 95-108) and Bednarek and Caple (2017: 17-21). First of all, migration is defined. Then the selection of data is presented and the newspaper article corpus as such and the image corpus outlined. Afterwards, common tendencies and characteristics of these images are discussed and selected examples analysed in detail. Finally, conclusions are drawn for the implied consequences that these images have for the interpretation of the newspaper articles.

Migration
Migration can be broadly defined as some form of movement of people that these accomplish over a certain distance and time. Such movements have decisive consequences for the migrating people themselves as well as for the receiving societies (see Mavroudi & Nagel 2016: 4). Different classifications have been put forward in order to show the manifold existing forms of movements that can be regarded as migration. Frequently used dichotomies are: -internal versus international migration: a spatial dimension, i.e. within or across countries; -temporary versus permanent migration: a temporal dimension, i.e. students, seasonal workers are examples of temporary migrants; -free/voluntary versus forced/involuntary migration: i.e. the migration decision as forced migrants are, for instance, people who are persecuted due to political or religious reasons; -individuals, groups or masses of migrants: i.e. the number of people involved.
Other migration types that show more complex forms of movements are, for instance, repeat migration, i.e. people migrate, return to their home country and then migrate again, or return migration, people returning to their home countries, or lifestyle migration, i.e. affluent people after their retirement migrate to another region that is better in terms of climate or for the people's health condition (Oltmer 2016: 17).
Such classifications as above (Treibel 2011: 20-22, 30, 172-173), however, simplify migration, which is highly complex in reality and hardly ever falls into one class only but more likely into more than one of them. They are nevertheless helpful in outlining the diversity of migration processes and are important for the distinction of legal terms and the naming of migrating people.
Terms that are used in this context are migrant, immigrant, emigrant, refugee, displaced person, asylum seeker and foreigner. Migrant is defined by the United Nations (UN) in the context of an international migrant, who is "any person who changes his or her country of usual residence [...]. Temporary travel abroad for purposes of recreation, holiday, business, medical treatment or religious pilgrimage does not entail a change in the country of usual residence" (UN DESA 1998: 9). The International Organization for Migration (IOM) defines a migrant more broadly: The movement of a person or a group of persons, either across an international border, or within a State. It is a population movement, encompassing any kind of movement of people, whatever its length, composition and causes; it includes migration of refugees, displaced persons, economic migrants, and persons moving for other purposes, including family reunification. (IOM 2011: 62-63) Historically, a migrant was only used for a person who made a free decision to move (IOM 2011: 61-62). In recent times, migration is rather used as a neutral or even as a cover term which includes refugees and other migrating people (Treibel 2011: 157).
An immigrant is someone that comes to a country, whereas an emigrant is leaving a country. For instance, immigrants are defined as "non-nationals [who] move into a country for the purpose of settlement" (IOM 2011: 49).
A refugee is defined in The Geneva Convention on Refugees of 1951 on the basis of individual persecution, and can be summarized as the circumstance that "a person cannot be returned to a country or territory in which his or her life or freedom would be threatened on account of race, religion, nationality, social affiliation or political opinion" (UN DESA 2017: 21). Displaced people have to move as consequence of a "forced removal" (IOM 2011: 29). They have to flee "for reasons other than those which would make him or her a refugee" (IOM 2011: 29).
Asylum seekers (also spelled as asylum-seekers) are "people seeking international protection and awaiting determination of their refugee status" (IOM 2017: 32). A foreigner is "[a] person belonging to, or owing an allegiance to, another State" (IOM 2011: 40).

Data selection and the corpus
As the aim is to examine migration in online newspaper articles of selected English-speaking countries, one online newspaper of each of the UK, the USA and Australia has been chosen. The decision of the newspaper chosen was based on different criteria: First, the newspaper with the highest readership was chosen, based on readership statistics for digital versions, if available. Here, Press Gazette (n.d.) for the UK, Media Miser (2016) for the USA, and Roy Morgan Research (2016) for Australia were used as sources. Second, online newspapers had to be accessible free-of-charge, guaranteeing an even higher readership than with newspapers that have installed a paywall. Third, only quality papers with a similar political position were chosen as a mixing of quality and popular newspapers or newspapers with different political alignments might lead to distorted results. In essence, the chosen newspapers are largely left-leaning.
As a result, the selected newspapers for the study are as follows: -The Guardian (GUA) for the UK -USA Today (UST) for the USA -The Sydney Morning Herald (SMH) for Australia News articles were then chosen from these three online newspapers and the time period was set to 1-31 August 2016. During this month, no particular events happened that could have lead to a one-sided reporting or might have influenced the coverage of migration in newspapers significantly. Furthermore, the corpus includes any article, whether a report or an editorial, for example, as news is here understood broadly (see also Bednarek & Caple 2017: 6). Newspaper articles were selected by using the search engine Google, as the homepages of the chosen newspapers all used different search facilities that could impact the search. The search was defined as that relevant news articles must include one of the following keywords within the above mentioned time period in the respective online newspapers: -migration/migrant/migrate -immigration/immigrant/immigrate -refugee/asylum seeker/asylum-seeker/asylum/asylee -unaccompanied minor -displaced person/displacement -foreigner/foreign-born/foreign national -non-national -protected person -alien These keywords were selected on the basis of terms used for migrating people, using predominantly glossaries from official organizations such as IOM (IOM 2011) or the UNHCR (UNHCR 2016).
The results returned 510 hits for The Guardian, 200 for USA Today, and 540 for The Sydney Morning Herald. From these, irrelevant results had to be sorted out manually, for instance, when migration was used in a biological sense, or when the same web page was returned twice. The corpus with the number of relevant articles and the number of words is given in Table 1. From these articles, all images were extracted. The numbers are given in Table 2.

Analysis of images
For the image analysis, all images in the corpus have first been classified. As was found, images frequently showed political figures, for example, Donald Trump, Theresa May, Peter Dutton and various ministers. As such, images used in articles about migration are highly politicised. In this context, Carvalho, Eatwell and Wunderlich (2015: 166-167) found that the British media, for example, play a key role in the politicisation of immigration. They are even more responsible for that than the government. Other images showed elite people such as sports or film stars, directors of organisations, buildings such as the White House, or non-migrant individuals who have no link to migration in any way. Some of these images might be considered to have an indirect relation to migration, but much more interesting are images that show migration more directly. As a result, a first distinction was made between images that are related to migration in a direct or rather indirect way. Some borderline cases such as protests against refugee and immigration policies are classified as direct relations as they depict migration as a controversial topic. If the link of the image to migration was not clear enough in the article, the image was counted to the nonmigratory part. An overview of the number of images depicting migration and not depicting migration is given in Table 3. Each of these two broad classes was further subcategorized. For non-migratory images, Table 4 gives an outline. Migratory images were examined in more detail, but a first classification into images depicting people or not depicting people is given in Table 5 (# for the number of images; % for percentage). As can be seen, the majority of migration-related images, i.e. about 80%, depict people.  Migration-related images were then grouped thematically (Figure 1). Most images here show detention centres or refugee camps. A considerable number of images depict protests against immigration policies or the conditions in refugee camps. Interestingly, some positive images occur as well, such as professional or educational success stories. Movements of migrants, especially of refugees, are depicted as they are on their way on roads and more frequently on the sea and in boats. Fences show migrants being excluded from their entry or from the freedom of movement in a country. Only few images depict migrant workers and migrants involved in crimes, terrorism and war. These categories already highlight that the number of images of refugees and asylum seekers predominate and that different types of migrants are not found in equal numbers in the corpus. can be interpreted as intimate or personal, a long shot shows the full body of a person from a distance and also implies emotional distance. A mid shot is in between these two types, showing the upper body of a person usually down as far as the waist. -Horizontal and vertical camera angle: Is the horizontal angle frontal or oblique? Is the image taken from a high, equal or low vertical angle? A frontal horizontal angle depicts the person not side on, as the oblique angle does. An oblique angle suggests that the viewer is an observer only. A frontal horizontal angle can imply that the reader becomes more involved with the person depicted. The vertical angle is linked to power relations. A high angle means that the reader is looking down on the people depicted and so the viewer is attributed more power, whereas a low vertical angle attributes more power to the people depicted as the reader has to look up towards them. -Demand or offer image: Is the person or are the people looking at the viewer or is the viewer simply an observer? If eye contact is suggested by the people depicted, the image is a demand image. An offer image shows the people as not looking at the viewer, i.e. they were not looking into the camera. -Balance framework: Is the image isolating or iterating? Which subtype of these two categories of composition is shown? The Balance Framework assumes that images are usually well composed and that they fall into several types. In essence, isolating images only show one salient element, iterating images show more than one element.
From all the thematic categories, some example images will now be analysed from the categories of borders and fences; boats, sea and beaches; detention/immigration centres and refugee camps; immigrants in general and migrant workers; and positive/success stories. General interesting tendencies within these categories will also be addressed.

Borders and fences
Here, borders between states are depicted. If it comes to refugees and asylum seekers, these borders are shown as unsurpassable, whereas for other migrants, these borders are not even clearly visible. Example 1 (below) and 2 illustrate that point. Example 1: Here we can see an image from a mid shot and in an oblique angle, and we see a man in dark-blue uniform as agent. He is placed on a diagonal axis with a tank, leading to an isolating, axial composition (see Caple 2013 for the distinctions of different categories). The tank links this scene to a military defence operation, although the man in the front is obviously not from the military. On the left is a fence with wire and the caption informs us that it is somewhere at the Hungarian border, presumably to the south. What is not shown are refugees or asylum seekers intending to cross the border. The man's responsibility is obviously securing the border and preventing certain migrants from entering the country. Example 2: If migrants other than refugees and asylum seekers intend to cross borders, they are rarely depicted with actual borders. More frequently, these borders are rather abstract and indicated as signs, as visible in this example. Furthermore, such migrants do not enter a country on foot, but much more often by plane. Only from the headline, not the caption, it becomes clear that the borders people cross here is used in an article describing EU nationals having immigrated to the UK. The image here is of an isolating, centred and single character. It is taken from a long shot, is at eye-level and has an oblique angle.
Migrating people in this category are not often depicted, but rather the border or fence or only an abstract symbol of that.

Boats, sea and beaches
Images in this category depict people on boats and ships, while they are on the sea or disembarking at beaches or ports. One typical example is given here: Example 3: In this image we find a dinghy filled with an uncountable number of people, most likely with an African origin, taken from a long shot. From the caption we know that it is somewhere in the Mediterranean. Differences among these people are largely neglected and invisible. Rather it is stressed that they are a homogeneous mass, showing a similar posture and a passive attitude. Therefore, the image is iterating, serializing and matching. The people are mostly looking at the viewer, thus, the image was taken from a frontal horizontal angle. The viewer is also attributed more power as we find a slightly high vertical angle. It is a demand image; the people are asking for help, although they do not wave and they do not seem to shout. However, no rescue boats are shown or anyone actually offering help.
Images such as example three showing an uncountable number of people from a long shot constitute only half of all images in this category. The other half are images that depict one or two individuals from a mid shot, who are singled out on a boat filled with many people. Thus, in this corpus, images depict people, mostly refugees and asylum seekers, not automatically as masses coming in boats, as was sometimes argued in previous research (e.g. Sauer 2016: 294; Martínez Lirola 2017: 141).

Detention/immigration centres and refugee camps
Being the largest category, it encompasses detention/immigration centres (69 images) and refugee camps (86). This difference is not always clear from the image itself, but the context helps here. However, there are not really differences in the way these images are composed. For images without people, Example 4 is illustrative. Figure 4: Entrance to the Verne immigration removal centre in Portland, Dorset. 'The windows look out onto more walls and barbed wire. In here you could be anywhere, and nowhere.' , Photograph: Alamy (GUA), https://www.theguardian.com/commentisfree/belief/2016/aug/11/despair-high-above-the-glittering-waters-of-weymouth-bay (date of last access: 11/01/2019) Example 4: This image looks like depicting a castle or a prison, given the thick walls and the tunnel as entrance. From the caption we know that it is an immigration removal centre in the UK. The image is taken from a long shot and is isolating, centred and single.
More images in this category, however, depict people. Typical images show them queuing up or waiting for something, thus, being largely passive. Fences are shown as well, especially people, individuals or uncountable numbers, looking out from behind the fences. Some images depict non-migrants. These are then helpers from organisations or teachers. In sum, a variety of age groups, numbers and different genders are shown. Most, however, are depicted in a passive way, as in Example 5: Example 5: This is a typical passive depiction of refugees and asylum seekers. The man sitting on the left is the most salient element, which is aligned diagonally to a few people placed on the right, thus being isolating and axial. The people are depicted as homogeneous mass, from a long shot, and thus the viewer being distanced. The image gives the impression of an unordered, gloomy situation, which is situated in Calais, as we are informed by the caption.

Immigrants in general and migrant workers
In contrast to refugees and asylum seekers, immigrants are often involved in some type of action, such as talking, listening, working. Nevertheless, images showing stereotypical behaviour or attributes occur as well, such as in Example 6. Example 6: In this image a man dressed in white garments is walking by Latino stores in New York. The image is of an isolating, axial composition with the man as the most salient element, taken from a mid shot. The man is a Muslim, as stated in the caption, and carries a so-called hajji, which men wear for a certain time after having been to Mecca. Though this might not be known to most readers, the image itself has no direct link to the news article and rather should be interpreted as illustration of immigration. It does that, however, in a rather stereotypical and restrictive manner.
Another group, migrant workers, are usually depicted as agents as well, normally as working. However, most images in the corpus show them in low-skilled jobs, such as as seasonal workers or as construction workers.

Positive/success stories
Not all images about migrants, especially about refugees, are neutral or negative. There are also a few images portraying successful people in educational or professional settings or as artists such as musicians, thus suggesting that they have become well integrated into and beneficial for society. These images never depict larger groups of people but often only one or two individuals. Example 7: This image depicts two men sitting at a table and one person holding a smartphone. The article text informs us that these two are refugees who have developed an app for other refugees. They are smiling and seemingly happy about their innovation, though stereotypes of refugees or foreigners are absent. The image composition is iterating, dividing and matching. It is taken from a mid shot and mostly frontal with the two looking at the viewer.

Conclusion
The aim was to collect online newspaper articles discussing issues about various forms of migration and to analyse their images. As shown, many forms of migration are not covered in the media; the picture of migration is thus quite unbalanced. As expected, most coverage is given to refugees and asylum seekers, migrants moving for other reasons are largely underrepresented, thus suggesting readers that migration mainly involves refugees and asylum seekers. This is consistent with what is usually expected from news, namely to report about newsworthy events and inform readers of issues relevant to them (see Bednarek & Caple 2017: 3). Still, it is discourse, not the events themselves that creates newsworthiness, as "events are mediated through language and image" (Bednarek & Caple 2012: 45). Nevertheless, the image corpus showed some variation in that a small variety of forms are at least addressed, such as migrant workers or EU nationals, apart from the high number of images depicting refugees/asylum seekers.
Within migration-related images, the largest category form detention/immigration centres and refugee camps, followed by protests, and positive/success stories. Within the protest category, migrants are hardly present. Refugees and asylum seekers are typically depicted behind fences or wires, being barred from entry to a country or from the freedom of movement within the host country. For other migrants, on the contrary, borders are hardly visible and depicted in an abstract manner.
Most images depict people (80%), but some buildings can give an intimidating impression, for example. If refugees and asylum seekers are depicted, they are shown as individuals, groups and also uncountable numbers of people and as mostly passive. If these people are portrayed positively, they are always shown as individuals. Migrant workers, on the other hand, are depicted as agentive, being involved in some action, though they are often subject to stereotypical representations due to their clothing or behaviour. This is problematic as generalization, suggesting that immigrants are very different from us, though the reality is far from that (as visible in the positive/success stories section, images can do without stereotypes).
In sum, this paper has outlined a comparatively large image corpus that was investigated with discourse analytic means. It has been shown that migration, even if we only consider images about refugees and asylum seekers, are indeed more varied than expected. As the corpus was based on left-leaning online newspapers, it would be worthwhile to examine if this holds for other newspapers as well.