CC BY
37EU immigration in the British press: How was immigration reported immediately prior to the EU referendum?
by Sawsan Askoul
Sawsan Askoul University of London sawsan.askoul@kcl.ac.uk
Published in Training, Language and Culture Vol 2 Issue 2 (2018) pp. 66-80 doi: 10.29366/2018tlc.2.2.5 Recommended citation format: Askoul, S. (2018). EU immigration in the British press: How was immigration reported immediately prior to the EU referendum? Training, Language and Culture, 2(2), 66-80. doi: 10.29366/2018tlc.2.2.5
This study examines the way in which European Union (EU) immigration was portrayed in the British press during the week immediately prior to the EU referendum of 23rd June, 2016. The research method is based on a combination of corpus linguistics and discourse analysis, with the aim of providing both quantitative and qualitative analysis of the issue. It begins by establishing the rationale behind the proposed topic, before outlining the theoretical and methodological framework of this corpus-based approach to investigating news discourse, which includes the analysis of frequency, concordances and collocates as they occur in three national-wide, online newspapers, namely the Daily Mail, the Guardian and the BBC. Further findings in relation to DA are then discussed, followed by concluding remarks and observations concerning the limitations of this approach and the tool utilised within it.
KEYWORDS: corpus linguistics, discourse analysis, immigration, European Union, media, collocates, concordances, referendum
S ®
This is an open access article distributed under the Creative Commons Attribution 4.0 International License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0)
It is a truism to say that immigration has been one of the most prominent topics in UK public debate over the past decade (Allen & Blinder, 2016). This reached its peak during the period of 'Brexit' - a journalese contraction of Britain exits the EU.
1. INTRODUCTION
editorials from around this time, justifying the various reasons why the British people were for or against their country being part of the EU. Among those who were in support of leaving, EU immigration was one of the most cited justifications.
Although Britain had been a member state since 1973, the British public had to decide on whether to leave or remain in the EU through voting in the EU referendum which took place on 23 June 2016 With a brief search of the internet, one can find
In accordance with 'the right of free movement', EU member states must allow all EU citizens to enter their countries and work without any
restrictions. As such, hundreds of thousands of Europeans have been permitted to live and work in
numerous online news reports, articles and
© Sawsan Askoul 2018
This content is licensed under a Creative Commons Attribution 4.0 International License
by Sawsan Askoul
'The BBC is very popular among British people, although it has been criticised for pro-government bias'
Britain. The Office for National Statistics estimates that there are more than 2 million EU nationals working in the UK.
Due to the increasing arrival and potential impact of EU immigrants to the UK as new states joined the Union, in particular those from 'A8' countries, the mass media has been covering this issue intently. The coverage has ranged from welcoming to fearful, with immigrants generally being viewed either as a much-needed workforce or a 'flood' of cheap labour, taking jobs away from the British population and causing an enormous strain on schools, health and housing services (Semotam, 2012, p. 2). Accordingly, this study aims to examine the discursive portrayal of EU immigration in the British press during the course of the week immediately prior to the EU referendum - from 16 to 22 June 2016.
The saliency and highly-contemporary nature of this issue lies at the heart of this study's rationale, which adopts elements of corpus linguistics in its approach. Indeed, as Stubbs (2004) observes, much of corpus linguistic research is motivated purely by curiosity. However, such research still
has a significant role to play within linguistics, even though it is aimed at description and interpretation of its findings, rather than critical evaluation (Baker & McEnery, 2015, p. 3).
2. MATERIALS AND METHODS
Corpus comes from a Latin term meaning 'body' (Baker & McEnery, 2015, p. 1). In its more modern sense, it can refer to a collection of texts of written or spoken language stored in electronic forms (O'Keeffe et al., 2007, p. 1). Corpus linguistics is a way of using computers to assist the study of language, since the sheer size of the corpus defies manual human analysis within any sensible timeframe (McEnery & Hardie, 2012, p. 2). However, this study of the language goes beyond traditional approaches to analysis, which might focus on linguistic units and structures such as morphemes, words, phrases, and grammatical classes, etc. (Biber et al., 1998, p. 2). With corpus linguistics, the focus is on 'real' language use of naturally occurring texts (Vessey, 2015, p. 1). As such, Baker and Gabrielatos (2008) believe that corpus linguistics is more a methodological than a theoretical approach, yet there is consensus that it provides both quantitative and qualitative analysis of actual patterns in use in naturally-occurring, authentic texts (Biber et al., 1998, p. 4).
As regards this study, the corpus was developed by adopting many of Sinclair's (2005) Basic Principles to provide a good practice framework. Firstly, the
Volume 2 Issue 2, 2018
rudn.tlcjournal.org
criterion of this corpus was set to include written, rather than spoken texts. Since written corpus can be gathered quickly using the internet (O'Keeffe et al., 2007, p. 5), news articles published on websites were compiled.
Secondly, balance was ensured by varying the selection to give a breadth of opinions, diverse political views, differing ideological positions and different target groups. Accordingly, the Daily Mail, the Guardian and the BBC were selected to ensure such range. According to a survey run by the internet-based market research firm, YouGov UK, the Daily Mail is seen as Britain's most right-wing newspaper. Statistics show its print title and its website reached an average 18.4 million individuals a month, through both its print and digital formats, from July 2015 to June 2016.
At the other end of the spectrum, the Guardian is seen as Britain's most left-wing newspaper. The BBC is very popular among British people, although it has been criticised for pro-government bias. Nonetheless, it is still a source of news being read world-wide.
Thirdly, the size of the corpus was carefully considered. Size is usually operationalised by the number of running words or 'tokens'. Heterogeneity is also important - essentially, it is how large the corpus is measured as 'word-types' or different parts of speech such as nouns, verbs,
adjectives, etc. (Stubbs, 2004, p. 113). This study is based on a corpus of nearly 20,000 words formed from 20 texts. O'Keeffe et al. (2007) consider any corpus below five million words as small, and certainly, when comparing it to other corpora such as the British National Corpus (BNC) - a 100-million-word collection of samples, this corpus is indeed small.
Yet, an additional measure of size should be considered, that of 'linguistic influence' or the size of the audience of the texts comprising the corpus. Since this corpus is based on online news articles, which tend to be freely accessible to a vast number of national and international readers, as the aforementioned Daily Mail stats affirm, this corpus can, therefore, be considered substantial in size.
Consequently, the corpus was built by searching each of the newspapers' online archives between the dates of 16 to 22 June 2016, using the search terms 'EU immigration'. So as to avoid any subjective selection and to present a systematic analysis, rather than presenting a covert polemic, all the articles containing 'EU immigration' in the titles only were chosen, instead of 'cherry-picking' the texts which could be seen as proving a preconceived point (Baker & McEnery, 2015, p. 5).
In order to make the corpus manageable, articles that contained 'EU immigration' in the body but
by Sawsan Askoul
not in title were discarded. The timeframe of the week immediately prior to the referendum was selected based on the thinking that news coverage would be at its peak during this time, and would potentially yield greater contributions and ultimately clearer findings concerning the topic of EU immigration with which to answer the question posed by this study.
3. STUDY AND RESULTS 3.1 Frequency
The data was stored in plain text format, comprising a single corpus of 20 separate files or sub-corpus consisting of 20 064 words in total. By using the developed version of AntConc 3.4.4m (Macintosh 0SX)2014 and referring to the work of Baker (2006), the analysis of the news discourse has been carried out to explore frequency, concordances and collocates.
Frequency is one of the most basic tools of the corpus linguistics. It is used to illuminate interesting findings and provide users with greater understanding of the use of a given word in certain contexts (Baker, 2006, p. 47). In this study, frequency analysis was used to investigate whether there was any variation in the news discourse or it was comparatively homogenous, whilst bearing in mind the different political orientations of each newspaper that comprised the data.
Although many corpus studies discard individual
'Although many corpus studies discard individual words as units of meaning, and instead investigate them as a part of phrasal units, words are still a good starting point, since the more frequent a word is, the more important it is to investigate'
words as units of meaning, and instead investigate them as a part of phrasal units, words are still a good starting point, since 'the more frequent a word is, the more important it is to investigate' (Stubbs, 2004, p. 115-116). This was achieved by obtaining the Word List or a list of all tokens and types, alongside their rank and frequency in the corpus. The initial results show that the corpus consists of 20064 tokens and 3099 types.
However, this raw frequency list reveals that the most frequent words in the corpus are 'functional words' - pronouns and determiners, rather than 'content words' - nouns, adjectives or verbs (Baker, 2006, p. 53). Usually, functional words are not subject to any further analysis. However, the abbreviation 'EU' (rank: 9, frequency: 217) and the content word 'immigration' (rank: 12, frequency: 165) appeared amongst the highest frequent words in the list (Table 1):
Volume 2 Issue 2, 2018
rudn.tlcjournal.org
Table 1
List of raw frequency
RANK FREQUENCY WORD
1 1299 the
2 661 to
3 528 of
4 453 in
5 432 and
6 395 a
7 311 that
8 281 is
9 217 EU
10 209 for
11 174 it
12 165 immigration
13 164 on
14 159 from
15 157 are
16 148 we
By clicking on 'Lemmas Word Form(s)' heading to the right of the word list in Table 1, other 'lexical lemmatised forms' that are directly connected to immigration were obtained. Lemmatised forms are defined as 'a set of lexical forms having the same stem and belonging to the same major word class' (Baker, 2006, p. 55). These included:
- migration (frequency: 83)
- immigrant (frequency: 9)
- immigrants (frequency: 53)
- migrant (frequency: 8)
- migrants (frequency: 54)
The high frequency of the abbreviation 'EU', 'immigration' and its lemmatised forms helps to give us a clearer idea about the main focus of the data.
A further consideration of the analysis was 'clusters'. The study of clusters can be understood as 'considering frequency beyond a single word' (Baker, 2006, p. 56). As such, it is useful to obtain a general sense of the portrayal of EU immigration in each of the newspaper texts. This was obtained by using the word list to search for 'immigration' in six-word clusters, as three or two
by Sawsan Askoul
word-clusters did not give any clear idea about the most common patterns in the corpus. The results showed that most clusters occurred with nearly the same frequency. Strikingly, one specific cluster that contains identical words and concentrated on exactly the same idea was common to all three newspapers:
The Daily Main (3 occurrences):
- immigration to the tens of thousands annually
- immigration to the tens of thousands and
- immigration to the tens of thousands, which
The Guardian (3 occurrences):
- immigration to the tens of thousands of
- immigration to the tens of thousands, which
- immigration to tens of thousands, featured
The BBC (1 occurrence):
- immigration to tens of thousands featured
Such identical clusters appear to suggest that the number of EU immigrants who have entered or may enter the UK was the one of the main concerns prior the vote. Moreover, other clusters and their repeat occurrences were likewise suggestive of the general attitude of each newspaper. For example, in the Daily Mail clusters such as: 'immigration system is a social disaster 'immigration, alongside the threat of Islamic terror'immigration and the migration crisis' were found. While the BBC: 'immigration threat or
opportunityand the Guardian: 'immigration should be controlled'immigration has hijacked the debate'immigration. People are scared
As it can be seen, immigration was evaluated negatively in those newspapers, although the negative evaluation tends to be more intensive in the Daily Mail in comparison to the other two newspapers. As such, readers are presented with the disadvantages of immigration and encouraged to make negative associations, especially were the word 'immigration' is clustered with words and phrases such as 'Islamic terror', 'crisis', 'threat', rather than say 'much-needed workforce', which actually does not appear in the corpus at all. Admittedly, the BBC appears to be more balanced by questioning whether immigration is 'threat or opportunity'.
The final stage in frequency analysis involves carrying out a dispersion plot. This provides visual representations of where 'immigration' is located in each text and, in turn, it allows us to compare the occurrences across multiple texts (Baker, 2006, p. 60.). 'Immigration' occurs at the beginning, in the middle or at the end of most texts. The most occurrences appeared throughout the Daily Mail text of 20 June with 29 occurrences, whilst it was only mentioned once in the middle of the Guardian text of 22 June, even though the article was published one day before the referendum. This pointed to the possibility that the Daily Mail had
Volume 2 Issue 2, 2018
rudn.tlcjournal.org
used the repetition as a technique to reinforce the notion that there was no possible way to reduce the immigration number without actually leaving the EU. This was then verified using the 'View File' feature to extract corresponding sentences from the Daily Mail text of 20 June, such as:
You may be surprised to read that I believe the Prime Minister has done everything in his power to control immigration. The problem is that as the leader of an EU member state, he does not have enough power to control immigration.
On many occasions, in public and private, I heard David Cameron set out his belief that precisely to protect this proud British heritage of welcoming people to our shores, immigration had to be controlled.
3.2 Concordances
Concordances are central to a corpus-based approach (McEnery & Hardie, 2012, p. 1). They are useful in conducting more qualitative forms of analysis than simply looking at frequent words/ phrases (Baker, 2006, p. 77). A concordance analysis is the most effective technique with which to carry out a close examination of the key words in the corpus.
By using the concordance function, concordance lines of the search word obtained. Each line contains the search word alongside a number of
context words which appear on either side of it in a word string. With the hope of obtaining different sorts of results, the separate searches were carried out for 'EU' in the Daily Mail, then the BBC, and finally the Guardian. By looking at the two words/ phrases occurring to the left of 'EU' in the 224 concordance lines, the results were as following:
The Daily Mail:
- number of EU
- Colchester, 77,000 EU
- as 270,000 EU
- were 178,000 EU
- at 188,000 EU
- three million EU
- two million EU
The BBC:
- at 188,000 EU
- three million EU
- number of EU
- figure for EU
The Guardian:
- number of EU
Predictably, the main focus of most articles is still on the number of immigrants that clusters analysis showed previously. This gives further indication that the number of EU immigrants was the main source of concern of all the articles. Furthermore, concordances give a deeper insight into how
by Sawsan Askoul
'This pointed to the possibility that the Daily Mail had used the repetition as a technique to reinforce the notion that there was no possible way to reduce the immigration number without actually leaving the EU'
immigrants were measured, with the exact numbers in most cases, in an attempt to provoke alarm over their growing number.
AntConc has a function called 'wildcard' which enables searching more than one word at a time (Baker, 2006, p. 75). For example, when adding an asterisk symbol (*) at the beginning or at the end of 'migrate', one can retrieve words such as migrants, migrant, migration etc. As for this analysis, * migrate and migrate* were fed into the search bar separately. Examples of concordance lines taken from the Daily Mail text of 22 June are as follows:
We control our OWN immigration: Soaring migrant numbers but independent Norway say
of independence saying it can cope with its soaring migration and run its own industries
when the situation is safer. It has seen soaring migration it is a member of the European Nevertheless, in order to understand the
effectiveness of such concordances in news discourse analysis, it is essential to understand the 'semantic prosody' they reveal. 'Semantic prosody' is a term used by Louw (1993) to mean a word that tends to collocate with certain other words, phrases or constructions in a particular environment, in a way that they render specific connotative and attitudinal meanings. Stubbs (1995) gives an example of the word 'cause' which tends to collocates with 'accident', 'cancer' and so on, and therefore is associated with negative semantic prosody. While 'provide' tends to collocate with 'food', 'help' and 'support', and is thus associated with positive semantic prosody (O'Keeffe et al., 2007, p. 14-15). Accordingly, the concordance lines of this corpus reveal that 'immigration' and its lemmatised forms co-occurred repeatedly with certain words and phrases, to give negative sematic prosody. They include:
- 'migrant crisis'
- 'wave of immigration' and 'flood of immigrants'
- 'soaring immigrant number', 'load of migrants' and 'increasing number of migrants''migrant influx' and 'queues of immigrants'
- 'EU immigrants were convicted of 700 crimes a week'
- 'undesirable migrants' and 'anti-immigrant'
- 'swarm of immigrants' and 'jungle, the migrant camp'
On the other hand, 'skilled immigrants' was found
Volume 2 Issue 2, 2018
rudn.tlcjournal.org
'The results show that 'cut' and 'control' are the strongest collocates associated with 'immigration' while 'future' and 'reducing' are the strongest collocates modifying 'immigration' in the whole corpus'
in the Guardian text of 20 June, which is an instance of positive sematic prosody, referring to EU immigrants through their contributions to the British society. However, the overarching context negates such positivity, and generates negative prosody overall, as shown in the extract below:
Not that the thorny issue of immigration was one that Cox ever sought to dodge. As recently as last week, she wrote a piece for the website Politics Home arguing that Britain had reaped many benefits from welcoming skilled immigrants. But she also acknowledged that, across the country, people faced everyday worries about job security, school places and CP appointments.
3.3 Collocates
Baker (2006) distinguishes between collocates -'when a word regularly appears near another word, and the relationship is statistically significant in some way'; and collocations - 'the phenomena
of certain words frequently occurring next to or near each other' (Baker, 2006, p. 95-96).
In the present discussion, this corpus study shall focus primarily on collocates. Echoing Allen and Blinder (2016), the collocates analysis focuses on verbs and modifiers that closely associate with 'immigration'. As the authors explain, 'modifiers are words that describe, characterise or intensify something' (Allen & Blinder, 2016, p. 7).
On AntConc, one can set 'collocates span' which determines how many words to the left or to the right of the search word to look at. For this analysis, it was set to one word to the left and then two words to the left. That made it possible to find phrasal verbs which could collocate with 'immigration'. AntConc also has a statistical measure which calculates how strong or weak a word collocates with the search word. The search was sorted by 'Stat', the collocation strength measure, that is based on 'Mutual Information' (MI), and the results show that 'cut' and 'control' are the strongest collocates associated with 'immigration' while 'future' and 'reducing' are the strongest collocates modifying 'immigration' in the whole corpus (Table 2).
Looking down the list, one can see that most of the collocates tend to produce negative semantic prosody, apart from a few exceptions such as 'defended immigration' and 'fairer immigration'.
by Sawsan Askoul
Table 2
The strongest collocates associated with 'immigration'
RANK FREQ FREQ(L) FREQ(R) STAT COLLOCA"
1 17 17 0 9.28823 about
2 4 4 0 9.20076 cut
3 2 2 0 8.20076 his
4 2 2 0 8.20076 future
5 11 11 0 8.07523 control
6 3 3 0 7.78572 reducing
7 3 3 0 7.78572 mass
8 1 1 0 7.20076 wider
'Looking down the list, one can see that most of the collocates tend to produce negative semantic prosody, apart from a few exceptions such as 'defended immigration' and 'fairer immigration'
In order to explore any possible day-to-day variation, a detailed analysis was carried out to search for all the verbs and modifiers which collocate with 'immigration', as they occurred in every article published on each day of the week immediately prior to the EU referendum. The findings were very interesting, revealing that the
Daily Mail tended to use collocates every day, but intensified such use just three days prior to the date of referendum, most of which conveyed negative evaluation.
The Guardian published an article on 22 June, containing no collocates, aside from functional words which have not been subject of any analysis.
The BBC, on the other hand, seemed to present more balanced arguments over the week, given that some of the occurring collocates were neutral-evaluative verbs such as 'comment on immigration', 'know about immigration' 'debate about immigration' and 'focus on immigration' (see Table 3).
Volume 2 Issue 2, 2018
rudn.tlcjournal.org
Table 3
Collocates having occurred in each article over the week
THE BBC THE DAILY MAIL THE GUARDIAN
Days Verbs Adjectives Verbs Adjectives Verbs Adjectives
16 June Control End Fairer Covered Could cut Decline in Net Control Pro-
17 June No article Want lower Fuelled by Increase No article
18 June Control Debate about Control Cut Solve More Net Control Pro-
19 June Worry about Comment on Reducing Push Increase Comment on Focus on Mass Increasing Debate on Relate Wider
20 June Control Controlling Meet Broken Control Low
Defended Control Net Talk about Public
Solve High Increase
Curb Think
Bring down
Decline in
21 June Lower Mass Cut Controlling Control Controlling
Control Rise in Reducing Focus on
Control Mass Know about
Relate High
increase
22 June Control Mass Reduce Mass - —
Controlling End Net
Cut Tough
Lower Firm
Clamp down Reducing
limit Soaring
Fuelled by
by Sawsan Askoul
3.4 Concluding analysis
The use of techniques associated with CL in carrying out DA is becoming increasingly popular (Baker et al., 2008, p. 275). DA is well-defined as the observations of the language patterns in use, as well as the context and circumstances associated with it, such as participants, situations, purposes and outcomes. Accordingly, discourse analysts tend to notice patterns systematically and to objectively produce, as far as possible, description, interpretations, explanations of their investigations (Trappes-Lomax, 2004, p. 133). The effectiveness of the synergy between CL and DA has been summed-up by Partington (2003), who explains that 'it helps to reveal examples of the phenomena that one has already noted. At the other extreme, it reveals patterns of use previously un-thought of. It can reinforce, refute or revise a researcher's intuition and show them why and how much their suspicions were grounded' (Partington, 2003, p. 85).
As for this study, the overall impression was that a negative portrayal of EU immigration is common to all three news sites - the Daily Mail, the Guardian and the BBC News. Immigrants were frequently quantified and referred to in terms of numbers, which has a rather dehumanising effect. Even the Guardian, which is seen as being a liberal broadsheet reflective of left-wing public opinion, tended to portray immigrants in this manner.
This study likewise found there to be a marked tendency towards immigrants being reported using metaphors of racist discourse. For instance, words such as 'jungle' were used to describe immigration camps, and immigrants themselves were referred to as a 'swarm' - words usually associated with animals and insects.
Van Dijk (1987) identifies four classes of topics within the racist discourse, and one of these can be observed in the news coverage which reported that immigrants are involved in negative acts that threatens the British society: 'EU immigrants were convicted in 700 crimes a week'.
As for 'wave of immigrants' and 'flood of immigrants', Van der Valk (2003) believes such metaphors symbolise the loss of control over immigration, and thus they can be seen to infer that the only way to control the process is to leave the EU. Other representations refer to the EU immigrants as 'queues' or 'influx', essentially, indistinguishable masses with no identity (Baker, 2006, p. 88). They are also labelled as 'undesired' and appear alongside words such as 'crisis', in a way that dehumanises them (Baker et al., 2008, p. 287).
Given these findings, it is worth considering how such negative representations of EU immigration in the press might actually translate into British public opinion.
Volume 2 Issue 2, 2018
rudn.tlcjournal.org
4. DISCUSSION
It is repeatedly quoted that media plays an essential role in our lives. The availability and popularity of social media and online news websites mean that press articles can be easily shared, reaching an enormous number of people. However, despite such wide availability, ordinary people generally have no direct effect on news content and they are not the major actors of news reports, rather this is the domain of the 'elite' such as journalists, scholars, writers and politicians, who have access to public discourses and communicative events (Van Dijk, 1995, p. 12). According to Baker and Gabrielatos (2008), these elites exercise their power by selection, acceptance or rejection of news content according to their agenda.
Consequently, they play an essential role in the reproduction of dominant knowledge and ideologies in society. However, newspapers also seek financial interests through sales, which are based on market forces. Therefore, they tend to report issues within the readers' concerns and aim to reflect their views. Readers usually pick up the newspapers that are in accordance with their perceptions and attitudes. Rejection, disbelief or criticism may be 'obstacles to the elites' access to the minds of the public through media' (Van Dijk, 1995, p. 13). For that reason, media tends to employ 'us and them' ideology, which Van Dijk (2006) refers to as 'an ideological square' which
focuses on the representations of two groups: ingroup's positive evaluation and outgroup's negative evaluations, emphasising 'our good things' and 'their bad things' and de-emphasising 'our bad things' and 'their good things' (Van Dijk, 2006, p. 734).
This model clearly demonstrates that media discourse has the ability to exert influence over the opinions and actions of its readership, who construct mental representations of the information they digest from new stories. In turn, these representations are influenced by what Van Dijk (1995) terms 'socially shared knowledge' or knowledge composed of prevailing beliefs and attitudes shared by the readers' social networks. When encountering news items, readers make use of this shared knowledge along with their own mental constructions in order to digest and make sense of what they are reading. Accordingly, Van Dijk observes that the media has a tendency to manipulate the structure of new items and their content so as to encourage particular mental representations and understandings among their readership.
Such manipulation might involve playing down or playing up the significance of certain information or individuals, giving greater or less coverage to particular people or particular topics, and of course deciding who or what makes it into the headlines.
by Sawsan Askoul
5. CONCLUSION
Although Corpus Linguistics helps to enable an understanding of complex and often ambiguous media representations (Baker & Gabrielatos, 2008, p. 33), it is not without its criticism. Whilst it employs computers to objectively count and reveal linguistic patterns, human input is still necessary to interpret those patterns. Such interpretations have the potential to be biased and subjective (Baker et al., 2008, p. 275), especially since analysis of CL tends to focus on specific words or phrases and disregard their context, which can give a distorted image of the truth.
Analysing the corpus of this study has not been conducted without limitations. The analysis is based on searching for certain content words and ignored functional words such as 'they', 'it' and 'this' that were repeatedly used to refer to immigrants and immigration. Expanding the search to include these pronouns and determiners may produce additional supportive findings or provide completely different results (Baker, 2006, p. 90). Moreover, unlike many other programs such as WordSmith, AntConc is lacking some tools that produce more thorough analysis. For example, it does not provide 'type/token ratio', which is useful in showing the complexity or specificity of corpus. Also, it does not have a frequency percentage column, which gives the overall proportion of how a particular word contributes to the whole corpus. Finally, due to the small size of the corpus, the
frequency and concordances of most key words/ phrases were very low.
References
Allen, W. L., & Blinder, S. (2016). A decade of immigration in the British press. Retrieved from http:// www.migrationobservatory.ox.ac.uk/resources/reports/ decade-immigration-british-press
Baker, P. (2006). Using corpora in discourse analysis. London: Continuum.
Baker, P., & Gabrielatos, C. (2008). Fleeing, sneaking,
flooding: A corpus analysis of discursive constructions of refugees and asylum seekers in the UK press, 1996-2005. Journal of English Linguistics, 36(1), 5-38.
Baker, P., Gabrielatos, C., Khosravinik, M., Krzyzanowski, M., McEnery, T., & Wodak, R. (2008). A useful methodological synergy? Combining critical discourse analysis and corpus linguistics to examine discourses of refugees and asylum seekers in the UK. Discourse & Society, 19(3), 273-306.
Baker, P., & McEnery, T. (Eds.). (2015). Introduction. In P. Baker, & T. McEnery (Eds.), Corpora and discourse studies: Integrating discourse and corpora (pp. 1-12). London: Routledge.
Biber, D., Conrad, S., & Reppen, R. (Eds.). (1998).
Introduction. In D. Biber, S. Conrad, & R. Reppen (Eds.), Corpus linguistics: Investigating language structure and use (pp. 1-12). Cambridge: Cambridge University Press.
Louw, B. (1993). Irony in the text or insincerity in the writer? The diagnostic potential of semantic prosodies. Text and Technology: In honour of John Sinclair, 240, 251.
McEnery, T., & Hardie, A. (2012). Corpus linguistics: Method, theory, and practice. Cambridge: Cambridge University Press.
O'Keeffe, A., McCarthy, M., & Carter, R. (2007). From corpus to classroom: Language use and language teaching. Cambridge: Cambridge University Press.
Partington, A. (2003). The linguistics of political argument:
Training, Language and Culture Volume 2 Issue 2, 2018
The spin-doctor and the wolf-pack at the White House. Routledge. Semotam, J. (2012). Public perception of A8 migrants: The discourse of the media and its impacts. Glasgow: BEMIS.
Sinclair, J. (2005). Corpus and text-basic principles.
Developing linguistic corpora: A guide to good practice, 1-16.
Stubbs, M. (1995). Collocations and semantic profiles: On the cause of the trouble with quantitative studies. Functions of language, 2(1), 23-55. Stubbs, M. (2004). Language corpora. In I. Davies, & C. Elder (Eds.), The handbook of applied linguistics (pp. 106-132). Blackwell Publishing. Trappes-Lomax, H. (2004). Discourse analysis. In I. Davies, &
doi: 10.29366/2018tlc.2.2.5 rudn.tlcjournal.org
C. Elder (Eds.), The handbook of applied linguistics (pp. 133-164). Blackwell Publishing. Van Der Valk, I. (2003). Right-wing parliamentary discourse on immigration in France. Discourse & Society, 14(3), 309-348.
Van Dijk, T. A. (1987). Communicating racism: Ethnic
prejudice in thought and talk. Sage Publications, Inc. Van Dijk, T. A. (1995). Power and the news media. In D.
Paletz (Ed.), Political communication and action (pp. 9-36). Cresskill NJ: Hampton Press. Van Dijk, T. A. (2006). Politics, ideology, and discourse. In K. Brown (Ed), The encyclopedia of language and linguistics (pp. 728-740). Oxford: Pergamon Press. Vessey, R. (2015). Corpus approaches to language ideology. Applied Linguistics, 38(3), 277-296.
