Article Review

JOURNAL OF CORPORA AND DISCOURSE STUDIES 2018, 2:X E-ISSN 2515-0251

ALESSIA TRANCHESE UNIVERSITY OF PORTSMOUTH

GETTING OFF BENEFITS OR ESCAPING POVERTY? USING CORPORA TO INVESTIGATE HOW THE PRESS REPRESENTED POVERTY DURING THE 2015 UK GENERAL ELECTION CAMPAIGN

CITATION Tranchese, A. (2019). Getting Off Benefits

or Escaping Poverty? Using corpora to investigate how the press represented

poverty during the 2015 UK general election campaign. Journal of Corpora and

Discourse Studies, 2:X–X

ABSTRACT This paper investigates the representation of poverty in nine national British newspapers during the 2015 UK general election campaign. A framework combining the qualitative approach of Critical Discourse Analysis with corpus-based techniques is used to address the following questions: which themes did the press foreground and background a) when poverty was mentioned explicitly, b) when it was referred to in terms of benefit claiming and c) when the two were mentioned together in the same article? The main claim of this study is that explicit discussion of poverty in the press was less frequent than the discussion of benefit claiming, and appeared to be mostly detached from the immediate context of the general election. This seemed to reflect a lack of attention paid to poverty in political discourse. Poverty was presented both as a reality without tangible causes and a common enemy that must be fought by an undefined entity. On the other hand, the welfare state (particularly benefit claiming) was presented as a burden that creates dependency; therefore, welfare reform and the reduction of expenditure on benefits — central to the general election campaign — were posed as a necessity for the reduction of debt. It is argued the separation of poverty and benefit claiming into two parallel debates, with only a small minority of articles discussing them in conjunction, allowed the press to sustain two incongruous messages at the same time: the need to both cut benefits and end poverty. It was only in the minority of articles that discussed poverty together with benefit claiming that a more visible counter-discourse surfaced, with more emphasis on the structural causes of poverty, the inadequateness of the benefit system and the positives of supporting people through the welfare state.

KEYWORDS corpus linguistics; critical discourse

analysis; media discourse; welfare; poverty

CONTACT Alessia Tranchese, School of Languages and Area Studies, Faculty of Humanities and Social Sciences, University of Portsmouth, Portsmouth PO1 2DZ, United Kingdom. [email protected]

DOI 10.18573/jcads.9

ORCID 0000-0002-3589-330X

ISSUE DOI 10.18573/jcads.v2i1

LICENSE © The authors. Available under the terms of the CC-BY 4.0 license

Manuscript accepted 2019-05-17

2 Journal of Corpora and Discourse Studies 2

Getting Off Benefits or Escaping Poverty? Using corpora to investigate how the press represented poverty during the 2015 UK general election campaign

Alessia Tranchese University of Portsmouth

1. Introduction There is no uncontentious definition of poverty, as its boundaries have shifted over time (Spicker, 2007). For example, in the 1970s, Townsend (1979) argued that poverty was more about the inability of people to participate actively in society than it was about a shortage of income. Echoing Townsend, Ferragina et al. (2013, p.37) argued that there exists a connection between income and participation, and that this connection should not be neglected in debates concerning appropriate ways of measuring poverty. More re- cently, in Section 7 of the Welfare Reform and Work Act (2016) the UK Conservative government replaced income-based measures of child poverty with measures of educa - tional attainment gaps and worklessness. However, in the absence of a universal defini- tion of poverty, more tangible manifestations have often been chosen as indicators. These include measures of income and material deprivation, such as low income, lack of access to essential goods and services, and benefit claiming (Spicker, 2002).

Among these three main indicators of poverty, it is particularly the accuracy of media coverage of benefit claiming that has often been denounced as problematic. For example, in 2011, at the height of the welfare reform, which made changes to the benefits offered within the British social security system, the Department for Work and Pensions Select Committee (2011, p. 3) defined media coverage of incapacity and disability benefits as ‘of - ten irresponsible and inaccurate’ with ‘pejorative language’ used by ‘some sections of the press […] when referring to benefit claimants’. Similarly, the Leveson Inquiry into the Culture, Practices and Ethics of the Press (2012, p. 684) found this coverage to be ‘biased’.

This study focuses on the public debate about poverty in the United Kingdom during the 2015 general election campaign (won by the Conservative party) to investigate whether it was characterised by a similarly biased coverage of benefit claiming and how this compared to explicit representations of poverty and poor people, that is to say, whether different ways of referring to poverty were associated with different representa - tions. To this end, it addresses the following questions: which representations did the press foreground and background a) when poverty was mentioned explicitly; b) when it was referred to in terms of benefit-claiming; and c) when the two were mentioned to - gether in the same article? The research presented in this paper builds on a larger project (Sippitt and Tranchese, 2015) that included print and online press, TV and radio, as well

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

3

as speeches by the leaders of the main parties in the UK. Due to word constraints, the dis - cussion in this paper will be limited to the press.

2. Poverty in media and political discourse Coverage of poverty in the British media has been found to be peripheral, with rare dis - cussions of its socio-economic causes and consequences, and with over-reporting of poverty as a problem of vulnerable people living mostly outside the UK (Chauhan and Foster, 2014; McKendrick et al., 2008). In contrast, the coverage of benefit claiming seems to revolve around ‘stigma stories’ that target specific characters, such as teenage or single mothers, and people with disabilities (Kelly, 1998; Pykett, 2014). Baker (cited in Baker and McEnery, 2015, p. 244) found that, over the ten years 2002–2012, the con- struction of benefit cheats and benefit culture had become particularly prominent in the right-leaning tabloid The Sun. Similarly, Baumberg et al. (2012, p. 86) identified a ‘consid- erable shift [in the media] towards describing claimants as ‘scroungers’, together with a consistently striking number of stories about fraud’. Another dominant theme in the pub- lic debate surrounding the welfare system seems to be the link between claiming benefits and the ability or willingness to work (Baumberg et al., 2012; Pykett, 2014).

Echoing the criticisms of media representations discussed above, these studies suggest that the coverage of benefit claiming is biased towards representations of the so-called ‘culture of welfare dependency’. The popularisation of this concept is strongly linked to a book by Charles Murray (1990, first published as an article in The Sunday Times Magazine), who, recalling the Victorian distinction between ‘deserving’ and ‘undeserving’ poor, controversially wrote about the ‘underclass’, a group of people whose behaviour was characterised by ‘illegitimacy, violent crime, and drop-out from the labour force’ (1990, p. 26). In Murray’s view, the underclass was not poor because of inequality, but be - cause of welfare provision that constituted an active disincentive against paid employ- ment and created a culture of welfare dependency by depriving people of the desire and discipline needed to work full-time. Murray’s notion of poverty related to welfare de- pendency as a life choice and alternative to work, though, is in contrast with a dominant body of literature that considers poverty a result of barriers created by structural inequal - ity and rejects the ideologies that generate and sustain negative understandings of and connections between poverty and welfare. In particular, it challenges the assumption that there exists a group of deserving and a group of undeserving poor, and that poor people must be punished to deter them from claiming benefits (Townsend, 1979; Spicker, 2002).

However, despite the fact that the evidence in support of Murray’s views is scant and does not stand up to scrutiny (MacDonald et al., 2014, p. 217), they seem to have been ac- cepted as the conventional wisdom not only by mainstream media narratives, but also in political discussions, with changes in political leadership not disrupting this discourse. For example, the Freud Report, commissioned by a Labour government, stated that ‘the

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

4 Journal of Corpora and Discourse Studies 1(1)

difficult heritage of the passive labour market policies of the 1970s is one of welfare de- pendence rather than self-reliance’ (Freud, 2007, p. 46). In a similar vein, the former Sec - retary of State for Work and Pensions and member of the Conservative Party, Ian Duncan-Smith, claimed that, because of the failure of previous governments to introduce welfare reform, ‘welfare dependency [had taken] root in communities up and down the country, breeding hopelessness and intergenerational poverty’ (Cabinet Office, 2010, p. 1).

The — very often interconnected — political and media discussion of poverty matters because, as suggested by Fairclough (1989, p. 54) ‘[the] effects of media power are cumu - lative, working through the repetition of particular ways of handling causality and agency, particular ways of positioning the reader, and so forth.’ Similarly, Stubbs (2001, p. 215) argued that ‘[r]epeated patterns show that evaluative meanings are not merely per - sonal and idiosyncratic, but widely shared in a discourse community. A word, phrase or construction may trigger a cultural stereotype’. Thus, by providing ‘hostile coverage [that is] stigmatising in itself’ (Baumberg et al. 2012, p. 86), for example by linking benefit claiming and anti-social behaviour like in TV programmes such as Channel 4’s Benefits Street (Smith, 2014), where benefit claimants are portrayed as criminals, media coverage can invite negative evaluations of poor people (Paterson et al., 2016, p. 212) and the per- ception that the welfare state supports ‘undesirable behaviour’ (Pykett, 2014, ¶2.5). Media representation of benefit claiming can also become a ‘morality play’ that influences people’s perception of the welfare state and benefits claimants specifically, shapes public debate and steers public policy (Kelly, 1998, p. 444), leading to the justification of anti- welfare arguments, as in the case of the UK government’s decision to subject people pre - viously judged unfit to work to a Work Capability Assessment (a test first introduced by New Labour in 2008) to re-evaluate their capability to work (Mooney and Neal, 2010). At a time of major political change like the general election, studying media coverage of wel- fare is particularly crucial, considering its potential to shape public perception of the wel- fare state, of its cost, and of the measures suggested by each political party to tackle this, thus, ultimately, to influence people’s vote.

3. Corpus building The time-frame for this study is 2015-03-30 to 2015-05-07, that is to say, the five weeks preceding the 2015 UK General Election. Data collection was conducted using the online database Factiva (Dow Jones, 2019) and the research included the daily, Sunday and on - line editions of the highest-circulation British national newspapers of various political al- legiances: the generally left-leaning The Guardian, The Independent and Daily Mirror, and the generally right-leaning The Times, The Daily Telegraph, Daily Mail, Daily Express, Daily Star and The Sun.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

5

It should be noted that a full comparison of newspapers is not the aim of this paper, as this intends to focus on the cumulative effect of dominant discourses, that is to say, ‘ways of representing the world or particular phenomena within it’ (Baker et al., 2013, p. 21). Comparisons between newspapers with different political allegiances will only be dis - cussed in relation to dominant trends highlighted by keyword and collocation analysis.

Right-leaning newspapers were over-represented in the corpus; while it would have been possible to remove some titles in order to have a balanced corpus, this would have resulted in a misrepresentation of news coverage. Preserving this imbalance may have skewed results in terms of which newspapers wrote the most about poverty or benefits, but it is nonetheless representative of the disparity that exists between left and right-lean- ing press in the current British press landscape. The fact that the British press is clearly skewed towards newspapers that traditionally express scepticism for the role of the state, including the provision of welfare is a significant element itself and should not be erased with a corpus composed of an equal proportion of right and left-leaning titles.

Poverty-and- Benefits corpus

(raw/norm.)

Poverty corpus

(raw/norm.) Benefits corpus

(raw/norm.) Total

The Sun (S) 25/10.5 61/25.7 151/63.7 237

The Guardian (G) 127/24 214/40.6 186/35.2 527

The Times (T) 45/14.8 103/33.8 156/51.3 304

The Daily Telegraph (DT) 46/14.3 96/29.9 179/55.7 321

The Independent (I) 55/16.2 115/34 168/49.7 338

The Daily Express (DE) 21/9.5 58/26.2 142/64.2 221

The Daily Mail (DMa) 51/11.7 110/25.3 273/62.9 434

The Daily Mirror (DMi) 31/15.4 52/25.8 118/58.7 201

The Daily Star (DS) 9/8.4 26/24.2 72/67.2 107

Total 410/124.8 835/265.5 1445/508.6 2690

Table 1: Number of articles per newspaper per corpus

A total of 2,690 articles were collected (duplicates were removed) and grouped into two corpora: the Poverty corpus (680,254 words) and the Benefits corpus (1,103,810 words). Table 1 contains the distribution of texts and shows their presence per newspaper; the number of articles was normalised per 10,000 for ease of comparison. The Poverty cor-

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

6 Journal of Corpora and Discourse Studies 1(1)

pus encompassed articles that contained at least one of the terms poverty, poor, poorer, poorest but not any of benefit, benefits, welfare, social security. The Benefits corpus contained articles with at least one of the terms in the second set but none in the former. These terms were chosen to separate explicit references to poverty from implicit ones made through one of its indicators. A third corpus of articles that con - tained both explicit and implicit mentions of poverty was collected and named Poverty- and-Benefits corpus (501,163 words). Each article in this corpus contained at least one word from each of the other two sets. Articles with irrelevant occurrences of these search terms, such as ‘the poor performance’ or ‘the country will benefit from’ were manually re- moved.

The Poverty-and-Benefits corpus was the smallest of the three and less than half the size of the Benefits corpus, with three newspapers (The Guardian, The Independent and Daily Mirror, all on the left of the political spectrum) accounting for almost half (44.5%) of the articles in this corpus. In the Benefits corpus, on the other hand, the majority of articles (72%) were in the right-leaning press, with The Guardian and the Daily Star con- taining respectively the smallest and the largest proportion of articles. While a difference in quantity does not necessarily translate into qualitative differences and, therefore, into a strong counter-discourse of welfare dependency in articles that referred to poverty expli - citly, this preliminary observation highlights a quantitative discrepancy that may indicate an ideological difference in how the two newspapers understand poverty as a social issue rather than as a fiscal one and may explain the different size and composition of each cor - pus in terms of representation of left and right-leaning newspapers.

An additional normative (not strongly contrasting) corpus (Rayson, 2002) was collec- ted using stratified week sampling in order to provide a representation of the overall mix of topics over the research time frame; the corpus contained all articles published over the artificial week of Monday 2015-03-30, Tuesday 2015-04-28, Wednesday 2015-04-22, Thursday 2015-04-02, Friday 2015-04-10, Saturday 2015-04-04 and Sunday 2015-05-03.

4. Methodology This study starts with a comparison between the Benefits corpus and the Poverty cor-

pus and the subsequent identification of keywords, i.e., words that are ‘statistically signi - ficantly more frequent’ (Baker et al., 2013, p.27) in one corpus than in another. This facil- itated the identification of themes that were more or less prevalent in one of the corpora. The statistical measure employed here to determine the significance of the difference is Average Reduced Frequency (ARF; Savický and Hlaváčová, 2002), available on the online corpus analysis tool Sketch Engine (Kilgarriff et al., 2014; Lexical Computing, 2019). ARF ‘discounts frequency for words with bursty distributions’ (Kilgarriff, 2009, p. 2), allowing researchers to combine statistical significance with information about word distribution, instead of focusing on simple frequency of occurrence which ‘may sometimes be mislead-

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

7

ing since [it does] not take into consideration the degree of dispersion of the relevant lin- guistic variable’ (Gries, 2008, p. 403). With this method ‘the whole corpus is considered as one sequence of words obtained by concatenation of all the texts forming the corpus’ (Savický and Hlaváčová, 2002, p. 217) and ARF is calculated by measuring the distance between occurrences of a word in this sequence and by assigning it a score that is close to its raw frequency, if it is distributed evenly within the corpus, or a smaller ARF score (usually one or little over one) if it is not distributed homogeneously (Kilgarriff, 2009, p. 2).

Consequently, since the ordering of the individual texts in the corpus will determine the sequence in which these words will appear, ARF scores will be affected by the organ - isation of texts within the corpus. In this study, texts were organised chronologically per newspaper. This means that articles covering the same stories for a certain number of consecutive days appeared next to each other, creating clusters of common or unique words. However, once the attention towards a certain story decreased, these terms either appeared less frequently or were discarded altogether. ARF made it possible to identify these occasionally very frequent words, because they had ‘a corrected frequency which is substantially smaller than the pure frequency’ (Savický and Hlaváčová, 2002, p. 217) and exclude them.

Similarly, ARF left out terms which would have been key simply because they were absent in the reference corpus, such as the proper names of benefit claimants. For ex- ample, a keyword list generated using log likelihood contained the name Mike Holpin, a benefit claimant who featured prominently in the news for the amount of money his family was claiming in benefits. However, his name was absent in the ARF keyword list because it almost exclusively appeared during two weeks in April and only in tabloids and right-leaning broadsheets; the ARF score excluded these occasional bursts of words in fa - vour of terms that were consistently frequent throughout the corpus. Although the study of bursty terms could be beneficial in understanding certain nuances of the poverty de- bate, for the purpose of this study these were discarded to avoid generalised claims due to their narrow distribution across the corpus.

When using ARF, a cut-off point for the statistical significance of keywords cannot be applied and all words with an ARF score close to their absolute frequencies should be analysed. However, analysing hundreds of keywords could result in too much informa - tion that could overwhelm the reader (Baker et al., 2013). As a result, only the top 100 lexical keywords were considered. These were then grouped together into semantic macro-categories in order to identify the topics that dominated the debate of poverty and benefit claiming when these were discussed separately. Keywords in each corpus were categorised using Wmatrix (Rayson, 2002), a web-based semantic tagger, in order to provide a systematic classification and reduce the risk of inconsistencies and bias. How - ever, in some cases, the meaningfulness of this categorisation had to be assessed manually

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

8 Journal of Corpora and Discourse Studies 1(1)

to ascertain its relevance. In some cases, for instance, the category applied by Wmatrix was modified on the basis of contextual information (for example, Sturgeon referred to the Scottish First Minister Nicola Sturgeon, and was classified under the ‘Politics’ cat - egory instead of ‘names of people’, as suggested by the software). In other cases, manual inspection was necessary to disambiguate certain terms. For example, child was not used referentially but often as a pre-modifier of benefits, indicating a specific type of welfare support. In some cases, however, keywords were considered too generic to be categorised in a useful way. For instance, system was used with welfare, benefits, tax and voting system, without a clear preference; thus, such terms were excluded from the list to avoid skewing the data towards a certain category when there was no vis - ible tendency. Therefore, despite the use of an automatic semantic tagger, the categorisa - tion and analysis of keywords was still partially subjective.

Generating keywords and grouping them into semantic categories were the starting point for the analysis and they provided a sense of the context for each corpus with an overview of dominant themes. Since it would not have been possible to analyse 100 keywords in great detail here, the terms poverty and benefit* — those whose absence, presence or co-occurrence distinguished one corpus from the other — were selected to study representations associated with different ways of referring to poverty (either dir- ectly or through one of its indicators). These terms, together with their collocates and concordances, were analysed to evidence more nuanced patterns and ‘get an impression of the ways that they contribute[d] towards discourses’ (Baker and McEnery, 2015, p. 250) of poverty and benefit claiming. Using WordSketches, a tool available on SketchEn - gine that provides a ‘summary of the word’s grammatical and collocational behaviour’, collocates of these terms were identified and analysed in context through concordance analysis. The minimum frequency of collocates was set to five, the SketchEngine default value. To facilitate reading, only random samples from the concordance — created using the SketchEngine sampling function — are provided.

Collocation analysis and close reading of concordance lines of poverty and bene- fit* was also carried out in the poverty-and-Benefits corpus to investigate whether their representation when they appeared in the same article remained similar to how they were represented when discussed separately. Thus, not only language choices that charac - terised each term in different contexts (that is, both in isolation and in conjunction) were identified and compared, but this was also intersected with the comparison between the behaviour of poverty and benefit* (and their collocates) across the three corpora.

5. Keyword analysis Based on both Wmatrix and the author’s manual categorisation (categories were kept as generic as possible to facilitate comparison across corpora), keywords in the Poverty cor -

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

9

pus were scattered across 22 semantic groups, while keywords in the Benefits corpus were concentrated in a smaller group of 14 semantic categories (Table 2).

Category Poverty keywords Benefits keywords

Employment jobs, work, wage

Government government, minister, parliament, public, state

Immigration immigration

Money rich, wealth, market borrowing, budget, cap, cash, cost, credits, cut, cuts, debt, deficit, economic, economy, fiscal, income, insurance, money, paid, pay, payments, savings, spending, tax, taxes

Places

Places: general city, countries, earth, local, planet, west

Places: national London Britain, British, country, Scotland, Scottish, UK

Places: international/developing world

American, developing, global, international, world, York

EU

Politics

Politics: general coalition, manifesto, plan, plans, pledge, policies, policy, politics

Politics: parties green conservative, conservatives, dem, democrats, dems, lib, liberal, parties, party, SNP, Tories, Tory, UKIP

Politics: voting campaign, election, general, poll, polls, referendum, seats, vote, voters

People

People: general boy, children, director, human, men, people, population, president, young

People: proper nouns Chris, Sarah

People: politics Balls, Cameron, chancellor, Clegg, David, Ed, leader, Miliband, MPs, Mr, Nick, Nicola, Nigel, Osborne, prime, Sturgeon

Time century, childhood, day, history, modern, recently

week, year, years, yesterday

Welfare allowance, benefit, benefits, child, disability, housing, nhs, pension, pensions, welfare

Society

Society: belonging to a group common, communities, community, social, society

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

10 Journal of Corpora and Discourse Studies 1(1)

Category Poverty keywords Benefits keywords

Society: charitable bodies charity, organisation

Society: education education, pupils, school, schools, teachers, teaching

Society: family and personal relationships

friend, friends, mother, parents

Society: religion church, religious

Entertainment art, book, character, drama, film, music, novel, stories, story, writer

Environmental issues change, climate, development, environment, sustainable

Resources energy, food, gas, water

Knowledge experience, learn

Life and death death, died, life, live, lives, living

Negative circumstances gap, inequality, issues, war

Positive circumstances good, improve, justice, love, opportunity, success, successful

Poverty poor, poorer, poorest, poverty

Research project, published, research, study, university

Table 2: Number of articles per newspaper per corpus

The attention paid to the categories of ‘Money’, particularly public expenditure, and ‘Politics’ (the two largest semantic macro-categories in the Benefits corpus based on relat - ive frequencies) shows that domestic issues were salient in the Benefits corpus compared to the Poverty corpus. While the focus on the election was not surprising, the presence of the ‘Politics’ semantic category (for example, Labour, SNP, Tory, manifesto) not only suggests that themes in this corpus included the discussion of the different political parties on the run-up to the general election and their approaches to welfare, but also that these themes were more salient in the Benefits corpus than in the Poverty corpus.

In order to test this hypothesis further, both corpora were compared against the normative corpus (see Section 3). Nine out of the top 20 keywords in the Benefits corpus compared against the ad-hoc corpus belonged to the ‘Politics’ macro-category, while the keyword list of the Poverty corpus against the ad-hoc corpus contained only one such keyword, thus strengthening the hypothesis that themes related to politics were more sa - lient in the Benefits corpus than in the Poverty corpus.

By contrast, the largest semantic macro-categories in the Poverty corpus were ‘Inter - national’ and ‘Environmental’ themes (for example, climate change, global, devel-

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

11

oping or world, which referred to a range of different places around the world, includ - ing developing countries), ‘Entertainment’ (such as drama, writer, book, film, mu- sic), ‘Society’ (for example, schools, community, church, charity), and ‘People’ (for example, boy, children, population).

On the basis of this preliminary inspection of keywords, it can be argued that poverty was presented as a social and geographically distant issue, happening at global level (for example, in the developing world), or as an abstract or idealised concept, with references to the arts (for example, films or books about poverty). By contrast, the portrayal of be - nefit claiming seemed to revolve around financial topics that placed the issue very much in the arena of the general election by presenting it as a conceptually more concrete mat - ter with potential repercussions in the national context. Although grouping keywords se - mantically was not a definitive process, with overlapping between categories, it was a helpful tool to highlight topics and story types that were prominent in one corpus but less salient in the other; this, in turn, provided context for the findings of the collocation and concordance analyses presented below.

6. Collocation and concordance analysis Having provided an overview of the main themes that characterised both corpora, this section focuses on the collocation and concordance analysis of poverty and benefit*. This section is divided into three parts: a) collocation and concordance analysis of poverty in the Poverty corpus, b) analysis of benefit* in the Benefits corpus, and c) analysis of poverty and benefit* in the poverty-and-benefits corpus.

6.1. Poverty

The WordSketch of poverty, presented in Table 3, showed that this term was fre- quently associated with (poverty and/or) unpleasant circumstances, such as homeless- ness, inequality, war and (ill) health. Moreover, one of the most frequent pre- modifiers of poverty was extreme.

Together with other adjectival pre-modifiers detected through concordance analysis (for example, soul-destroying, serious, lethal, destructive, unacceptable; see Concordance 1), this was part of a group of pre-modifiers that cumulatively accentu- ated the negativity of poverty and seemed to suggest that poverty is scalable (does the phrase ‘serious poverty’ suggest that there are non-serious types?). The negativity of poverty was reiterated by its use as subject of BLIGHT in relation to lives, people, com- munities and homes. This presented poverty as an active entity that destroys and causes pain and suffering, rather than a consequence of actions and policies.

BLIGHT only appeared in left-leaning papers and tended to occur in the context of the arts with references, for example, to films about poverty, while adjectives that emphas - ised the negativity of poverty tended to appear in relation to international poverty (par -

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

12 Journal of Corpora and Discourse Studies 1(1)

ticularly in the right-leaning press). This echoed the findings of the semantic grouping of keywords. However, while the keyword analysis did not highlight ‘National Context’ as a salient category in the Poverty corpus, the concordance analysis showed that, in a few cases and especially in relation to particularly vulnerable groups such as children and young people, poverty was connected in the left and right-leaning press to the British context (see Concordance 1, lines 2, 7 and 9).

modifiers of poverty fq score nouns and verbs modified by poverty

fq score verbs with poverty as object

fq score

204 26.4 173 22.4 189 24.4

extreme 39 12.1 line 22 11.3 tackle 26 11.4

fuel 18 10.7 chastity 8 10.5 eradicate 15 11.2

child 14 10.5 inequality 7 9.98 reduce 14 10.3

food 8 9.26 eradication 5 9.83 fight 9 10.3

energy 6 9 pay 6 9.81 escape 8 10.3

reduction 5 9.61 end 11 10.2

project 5 9.01 alleviate 6 9.94

grind 5 9.71

address 5 9.34

be 14 6.14

verbs with poverty as subject

fq score poverty and/or … fq score prepositional phrases fq score

69 8.91 213 27.5 338

blight 6 11.3 inequality 15 10.9 … of "%w" 95 12.3

be 31 6.78 homelessness 13 10.9 … in "%w" 74 9.56

chastity 11 10.6 %w in … 35 4.52

war 7 9.81 … to "%w" 24 3.1

destitution 5 9.55 … from "%w" 13 1.68

unemployment 5 9.45 … by "%w" 12 1.55

health 5 8.88 … into "%w" 11 1.42

… on "%w" 11 1.42

%w of … 9 1.16

… with "%w" 8 1.03

… as "%w" 7 0.9

%w by … 6 0.78

%w for … 5 0.65

%w as … 5 0.65

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

13

modifiers of poverty fq score nouns and verbs modified by poverty

fq score verbs with poverty as object

fq score

… against "%w" 5 0.65

Table 3: WordSketch of poverty in the Poverty corpus, minimum frequency=5

Most verbs patterning strongly with poverty as object had meanings associated with the need to end it (e.g., TACKLE, ALLEVIATE, ERADICATE, REDUCE, FIGHT, ESCAPE; see Con- cordance 2). Here the focus on poverty as a distant and abstract issue was even more ac - centuated, with references to science (line 4), world migration (lines 6 and 7), the arts (line 1), and world poverty (lines 2, 3, 5 and 8). The pattern appeared in newspapers of different political allegiances, with a majority of articles (51 out of 81, approximately 63%) in the left-leaning press.

Source

1 looked ground down after a century of desperate poverty . The dismal state of the USSR's economy, DMa

2 that there are students are living in serious poverty is completely unacceptable. Many parents I

3 . Desperate to escape war, famine and lethal poverty , mums and dads are risking everything in the DT

4 , much like Pacquiao, was born into acute poverty and hardship. None of this offers mitigation DT

5 HARD-PRESSED consumers are living in abject poverty as they can't afford their energy bills, a S

6 soulful group portrait of a community blighted by poverty , and a triumph-of-the-human-spirit story I

7 young people are being blighted and degraded by POVERTY and homelessness. “POVERTY and homelessness take DT

8 dazzling talent whose early life was blighted by poverty , and whose years of celebrity were scarred by G

9 Chris Keates said: "Lives are being blighted by poverty . Pupils cannot concentrate because they are tired DMi

Concordance 1: Adjectival pre-modifiers of poverty in the Poverty corpus

Source

1 applicants per post. As the quest to end poverty remains in the limelight, young people G

2 that there are plenty of ways to alleviate poverty and empower women; the job ahead is to G

3 service to those struggling every day to escape poverty and homelessness. 'We are thinking big to address DMa

4 through social policies aimed at reducing family poverty could change the trajectory of brain development and I

5 development goals will aim to eradicate poverty by 2030 but our current economic model, G

6 to suggest it could include people simply escaping poverty added Mr Farage. DE

7 have died from countries such as Libya in order to escape poverty and humanitarian disasters I

8 economy. Governments can't hope to tackle poverty and meet other development goals without G

Concordance 2: Verbs with poverty as object in the Poverty corpus

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

14 Journal of Corpora and Discourse Studies 1(1)

In the British National Corpus (BNCweb, 2018; Hoffmann et al., 2008), these verbs often co-occur with negative circumstances like diseases (for example, to eradicate an infec- tion, tumour or disease), conflicts and risky situations (such as to fight a battle, war, or blaze) or captivity (for example, to escape death or to escape from prison). Similarly, TACKLE, ALLEVIATE and REDUCE are used in contexts such as tackling prob- lems, issues, crisis or alleviating suffering, symptoms, anxiety and reducing costs or risks. By associating poverty with these verbs, their negative semantic pros- ody — the ‘consistent aura of meaning with which a form is imbued by its collocates’ (Louw, 1993, p. 157) — is projected onto poverty, which, in turn, inherits the same at- tributes of diseases, conflicts or, more generally, enemies. As shown below, the personi - fication of poverty as a negative entity or enemy was a trend in the Poverty corpus.

ESCAPE was the only verb in the list of verbs with poverty as object that could imply the active role of poor people to end poverty. However, ESCAPE also implies the presence of someone or something that traps people and that, consequently, makes it hard for them to free themselves, thus taking away their responsibility for being poor. Similarly, the second most frequent prepositional phrase featuring the term, in poverty, provided further evidence that poverty was associated with a sense of imprisonment, through expressions such as living in poverty, people in poverty or being in poverty, suggesting that poverty is a state (Kress, 1994, p. 29) or condition people live in (not on nor with), a something that contains or traps them.

The other verbs with poverty as object referred to some form of external help or in- tervention to eliminate poverty, again placing the responsibility for getting out of poverty away from poor people themselves. However, those supposed to help end poverty or free those in poverty tended to not be presented as the grammatical subjects of such verbs, but appeared in infinitive or participial forms, or in the future tense, present - ing the end of poverty as a project, a promise or an aim for the future (see Concordance 2). Agents of help to end poverty were often vague or abstract entities, such as Britain, awareness and vision of fairness and equality, but also money, organisa- tions or leaders of foreign countries thus strengthening the sense that when poverty was mentioned explicitly it was often as an abstract or geographically distant matter.

Some of the verbal collocates with poverty as object were used metaphorically (for ex - ample, FIGHT, ERADICATE, FREEZE, SLASH, CUT). In their literal meaning, FIGHT or ERADICATE are respectively used in the context of war and removing roots (Miller, 1995; Princeton University, 2010). While metaphors are particularly useful conceptual devices that enable complex phenomena to be represented through more familiar imagery (Char- teris-Black, 2006), they can also oversimplify them. For example, The Guardian reported (emphasis mine):

(1) Financial literacy is […] an important tool in the fight against inequality and poverty

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

15

Here the metaphor was both grammatical (nominalisation, cf. Halliday, 1985), with a process expressed through a nominal form that facilitates vagueness by removing the need to specify who is fighting what, and conceptual, where a fight requires an opposi- tion between two parts: the fought against (normally a negative entity, here poverty) and the fighter (normally a positive entity, here an undefined entity). This opposition led to the personification of poverty as a common enemy, presenting everyone else — including those very institutions that fail to address poverty or even profit from it (such as credit institutions) — as (potential) parts of the ‘good’ side. As Hart (2014, p. 160) has put it, in a crisis situation, ‘reference to any culpable causers’ can be obscured through metaphor, ‘thereby concealing the role that particular institutions and political systems played in bringing it about’. A paradoxical representation ensued: poverty became both a state and the cause of that state, while the social structures that sustain it, the concrete measure that must be taken, and, in some cases, those who need to take them, were amalgamated in an extremely vague, but positive nonetheless, concept.

The lack of focus on the causes of poverty was also visible elsewhere throughout the corpus: for example, none of the verbs used with poverty as object referred to causes of poverty (e.g., CAUSE) and, while the most frequent prepositional phrase with poverty was of poverty, this occurred in expressions such as cycle of poverty or level of poverty but never in cause(s) of poverty. Moreover, a group of pre-modifiers of poverty were used in noun phrases such as fuel, food, energy and child-poverty. As suggested by Bhatia (1992, p. 202) noun phrases can have consequences in terms of discourse and ideology by obscuring the relationships between its constituents (here head noun and modifier). Furthermore, according to Jeffries (2010, p. 16) ‘the main ideological importance of a noun phrase is its ability to package ideas or information which are not fundamentally about entities, but which are really a description of a process, event or ac- tion’. Thus, like with metaphors, using poverty in noun phrases led to the obscuration of the links between institutional actions and policies — like government’s cuts to winter fuel allowance — and their outcomes, with the latter transformed into a vague container word, poverty, that was no longer clearly linked to the former but became an independ - ent entity. It can be argued that metaphors and noun phrases rendered the discussion of the causes of poverty, of what exactly needs tackling, and by whom, unnecessary, with the effect of obfuscating institutional responsibility under the impression of clarity and, consequently, supporting — rather than challenging — the status quo.

Overall, poverty was often discussed in emotional terms, as something negative, an unavoidable (rather than human-made) ‘disease’ that must be ended or collectively ‘fought’. If, on the one hand, it was constructed as an abstract and independent entity without causes, on the other it was also personified as a clear, identifiable, enemy blamed for people’s misery. What was absent was a strong discourse of poverty linked to unem - ployment, poverty wages, poverty pays, or in-work poverty.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

16 Journal of Corpora and Discourse Studies 1(1)

6.2. Benefits

Most pre-modifiers of benefit* (Table 4 — Appendix) were terms that referred to be- nefit types, particularly child benefits — another noun phrase where, however, the head noun refers to a less vague concept than poverty — possibly reflecting the significant focus on reductions of payments for parents during the election campaign. The other types of benefits could be divided into those that were people-oriented and not means- tested (child, disability and incapacity) and those that were employment-ori- ented and (partially or fully) means-tested (housing, in-work, pensioner, working- age, unemployment, pension, retirement, tax, out-of-work) (Government Di- gital Service, 2019).

Most of the verbs that co-occurred with benefit* as object could be grouped into verbs that referred to the perspective of a) those who administer benefits ( CUT, LIMIT, RESTRICT, DENY, TAX, CAP, LOSE, SLASH, STOP, ABOLISH, REDUCE, CLAW, FREEZE, PAY and REMOVE) or b) those who are at the receiving end of these policies (CLAIM, RECEIVE, GET, OBTAIN, ACCESS). Moreover, most of the verbs in the former category (apart from PAY) could be further divided into those that involved a drastic or permanent loss ( CUT, LOSE, SLASH, STOP, ABOLISH, CLAW, REMOVE) and those that indicated a restriction or reduction, without necessarily implying a complete loss (LIMIT, RESTRICT, TAX, CAP, REDUCE, FREEZE, DENY). What these verbs have in common is that they all suggest a downward trend in welfare payments, a suggestion existing in contrast with the emphasis on ending poverty emerged in the Poverty corpus.

Most verbs that indicated a drastic loss co-occurred with child benefit* (Table 5 — Appendix). CUT, for example, tended to be used in the context of the Conservative party cutting child benefit; this was presented as something undesirable or a threat, as shown by phrases that expressed the need to hide (for example, ‘secret’ plan to slash ) or negate:

(2) Mr Cameron […] did not want or had no plans to cut child benefit

In contrast, verbs that referred to a reduction did not tend to have a strong preference for a specific benefit type, but appeared particularly in relation to migrants and unemploy- ment benefits:

(3) EU job hunters would be denied unemployment benefits

(4) Restrict benefits for migrants […] to those who have been paying tax […] for five years

Verbs that referred to those who received state support — rather than those who provided it (in terms of welfare policy) — stressed the agency of benefit claimants and were also not typically used with child benefits. The strong emphasis on an (un- desirable) drastic loss of support for parents — not seen with other types of benefits —

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

17

may indicate that child benefits were seen as an automatic right of those with children, unlike non-universal benefits for which entitlement can be challenged more easily and loss may not be perceived as necessarily negative because these are dependent on income levels and, therefore, potentially more controversial. The fact that these verbs that em - phasised a request and the action of taking something, like taxpayer money from the state) seemed to appear in the context of fraudulent activities (cf. Concordance 3) seems to support this theory. This pattern was particularly visible in right-leaning tabloids which focused on cases of people who received support ‘undeservingly’ because ‘not en- titled’ to it, as suggested by modifiers such as fraudulently and dishonestly.

Source

1 the offence of two counts of obtaining benefits by deception, and will appear at Belmont DMa

2 Officers found that she had been fraudulently claiming benefits and had allowed others to use her name DMa

3 than material goods." Mr Holpin has been receiving benefits for 13 years and admits his […] handouts fund his 20-a-d DE

4 admitted fraudulently obtaining incapacity benefit and Jobseekers and employment allowances. DS

5 Cowan, who pled guilty to obtaining benefits to which he wasn't entitled between May 2011 and S

Concordance 3: Verbal collocates of benefits referring to welfare in the Benefits corpus

The theme of benefit fraud (Lundström, 2013) was also visible in the list of nouns and verbs modified by benefit*, where several terms (fraud, scrounger, cheat, tour- ism and scam) indicated dishonest activities. Claimant and system (Concordance 4) also appeared in the context of fraud, exploitation, and (extreme) generosity of the wel- fare system that was presented as a burden and, therefore, in need of change. The theme of benefit fraud was not completely absent in broadsheets and left-leaning papers, as shown in line 8 (Concordance 4), where it was embedded in a quote by a Labour spokes - person published by The Guardian.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

18 Journal of Corpora and Discourse Studies 1(1)

Source

1 years. Related articles Water big cheat: Benefit claimant who said she couldn't walk was DE

2 shedding a load of weight. The six-times wed benefits claimant outraged the nation when he appeared S

3 of 1,831 mothers found 52 per cent believe the benefits system is too generous for those who refuse to DE

4 has also pledged to crack down on abuse of the benefits system by EU nationals. In November 2013, he DMa

5 BRITAIN'S soft-touch benefits system has been blamed squarely for ruining the DS

6 the Taxpayers' Alliance has now warned that the benefit system should not become a 'comfort blanket' DMa

7 of more budget cuts. The Government cannot bring benefit fraud to an end without the help of DMi

8 security. We’ll save £1bn by cutting housing benefit fraud and overpayments and control housing benefi G

Concordance 4: Occurrences of benefits together with claimant*, system* or fraud* in the Benefits corpus

Finally, the verbs with particle off and benefit* as object (in live off benefits, be off benefits or get off benefits) showed how the strong association between receiving state benefits and dependency is entrenched in language use. In the BNC (BNCweb, 2018), liv* off tends to have a semantic prosody that indicates a bur- den or exploitation, as well as dependency on some form of criminal or immoral activity (Concordance 5). In the case of benefits, living off benefits represents the immoral (or sometimes criminal) ‘exploitation’ of the system, while being off or getting off benefits means the end of claimants’ dependency (like one gets off or is off addictive substances, such as alcohol or drugs), consequently unburdening the taxpaying public, the welfare system, and the government.

1 want money. What for? We could get by, living off parents, friends or the State. And

2 not right. She doesn't do anything. She lives off his wages.' 'Does he lose his

3 like that you can start getting done with living off immoral earnings, brothel-keeping

4 , in May. He added: 'She has been living off prostitution and is a heroin addict

5 accomplices are determined characters: they live off crime, and take as much as they can

Concordance 5: Occurrences of liv* off in the BNC

The analysis of the Poverty corpus and Benefits corpus showed that poverty and benefit claiming were represented as two fundamentally unrelated phenomena. On the one hand, poverty was mostly removed from the UK context and was discussed in general terms as a global issue, or the subject of books, films or music. It was also represented as an ex - tremely negative condition with particularly intense terms such as abject or BLIGHT. On the other, the reduction of welfare (with the exception of child benefits) was presented

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

19

positively, as unburdening of the state, and was linked to the individual’s (un)willingness to work, rather than to structural inequality. Thus, while there appeared to be an almost romanticised idea of poverty, as an entity without specific (institutional) causes, the other focused on individual responsibility.

This discrepancy echoes the ideological distinction made by Murray (1990, p. 24) between the poor who ‘simply lived with low incomes’ and the ‘feckless’ poor who chose not to work. Both major political parties expressed similar views during the election, with David Cameron — the then leader of the Conservative party — evoking Murray’s view of benefit claiming as ‘life choice’ — ‘when you’re getting up at the crack of dawn to get to work, you want to know that you’re not putting in all those hours to pay for someone else who chooses not to take a job’ (Conservative Party, 2015) and Ed Miliband — the then leader of the Labour party — emphasising entitlement to welfare — ‘fair rules means that entitlement to benefits needs to be earned’ (LabourList, 2014). The latter re- flects a broader shift in the Labour party towards a more populist discourse started with New Labour and characterized by the blurring of the boundaries between traditionally left and right-wing ideologies (Fairclough, 2000). With the introduction of a more au- thoritarian discourse under Toby Blair’s leadership, New Labour shifted towards ‘indi - vidualist discourse which stands in contrast with the traditional collectivism of the centre-left and the left’ (Fairclough, 2000, p. 40). In this sense, Miliband's speech evoked the triad ‘rights, responsibilities and duties’ mentioned by Blair in an article written for the Daily Mail in 1998 in which he used the expression ‘something for something’ that echoed the right-wing reference to benefit claimants as ‘spongers’ with ‘something for nothing’ (q. in Fairclough, 2000, p. 39).

6.3. Poverty and benefits

This section focuses on the ways in which poverty and benefit* were discussed when they appeared in the same article. This was the smallest corpus of the three (see Section 3). Given that, as discussed above, opposing political parties seemed to draw on similar ideas during the 2015 general election and considering the frequency with which the press reports the voice of institutional sources (Richardson, 2007), it may be hypothesised that the less common connection between benefits and poverty reflected the homogen- eity of the ideology surrounding welfare and poverty in political discourse, an ideology according to which benefit claimants cannot be legitimately considered poor.

The WordSketch of poverty (Table 6 — Appendix) in this corpus showed that the emotional representation of poverty found in the Poverty corpus, achieved through an abundance of pre-modifiers that highlighted its gravity, was less accentuated here, with no adjectives referring to its negativity. Instead, poverty was associated with unemploy- ment (or lack thereof) in various ways (for example, in the recurring collocations in- work poverty, poverty wage, poverty pay, poverty and unemployment), sug- gesting that work and money were particularly salient topics. Additionally, these colloc-

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

20 Journal of Corpora and Discourse Studies 1(1)

ates occurred mostly in the context of poverty in Britain, such as in relation to zero-hour contracts and food banks, with a reduced focus on poverty as a global issue (a particularly salient theme in the Poverty corpus). Porn as a post-modifier of poverty also strengthened the sense that poverty was anchored to the British context, since poverty porn consistently referred to the depiction of poverty in the British media, particularly in the TV programme Benefits Street. While more accentuated in the left-leaning press (per- haps because it is over-represented here; see Section 3), these patterns appeared in all newspapers, as shown in Concordance 6.

These trends were in contrast with those in the Poverty corpus, in which more ab - stract terms (for example, inequality, development, justice), terms related to fic- tion (such as film, novel, book, music) or negative adjectives were more prominent. Here, instead, poverty appeared to be less abstract and geographically distant, understood in less absolute terms and linked more clearly to context-specific social factors like finan - cial and employment issues. In this sense, this representation was more similar to that of benefit claiming in the Benefits corpus.

Source

1 ? Angela or the supermarket that pays her a poverty wage? And is this the Tories' great sunshine DMi

2 the "tip of the iceberg" of a food poverty crisis. Yesterday Labour accused Iain Duncan Smi I

3 were the “tip of the iceberg” of food poverty in the UK, while doctors said the inability G

4 wages low – in effect, a state subsidy for poverty pay. As the Office for Budget Responsibility puts DT

5 government's priority should be "to move from poverty pay to living pay". "Employers will have to T

6 in our society are plunged being deeper into poverty , where in-work poverty is on the rise. ' DMa

7 , putting food on tables, helping people out of poverty ." Average pay in the UK is up 1.8 per S

8 Britain of food banks, zero-hours, in-work poverty , housing crisis, job insecurity and young people G

Concordance 6: Occurrences of poverty together with wage, pay, work or food in the Poverty-and-Benefits corpus

The modifiers of and verbs with benefit* as object in the poverty-and-Benefits corpus (Table 7 — Appendix) coincided with those in the Benefits corpus; however, they ap - peared in different contexts. Verbs implying a loss (apart from ABOLISH) did not tend to co-occur exclusively with child benefits (for example, child benefits co-occurred with CUT only 5 out of 19 times), while, among the verbs implying a reduction, RESTRICT and LIMIT strongly (6 out of 8 occurrences) co-occurred with child benefit*. The negativity associated with cuts to child benefits visible in the Benefits corpus (see Section 6.2.) seemed less stressed here, given that, arguably, by specifying the nature of the reduc - tion, i.e., limitations or restrictions (instead of cuts) the press presented the issue in more concrete and moderate terms.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

21

This hypothesis was supported by the fact that, for example, when child benefit* occurred with RESTRICT in the poverty-and-Benefits corpus (Concordance 7), there was a clear specification of how many children benefit payments would have been limited to and/or for how long (two children, for two years), while, when child benefit* ap- peared as object of CUT in the Benefits corpus (Concordance 8), there was no specifica - tion of the timeframe or number of children, arguably implying more drastic measures that could lend themselves to a more sensationalistic and alarmist reporting. As shown in Concordance 8, this pattern appeared in both left and right-leaning papers.

One of the most remarkable differences between the Benefits corpus and the poverty- and-Benefits corpus was the almost total absence in the latter of terms that referred to fraudulent activities, a salient topic in the Benefits corpus (Section 6.2). The only excep - tion was fraud (Table 7 — Appendix); however, benefit fraud was used mostly (in 5 out of 6 occurrences) in left-leaning newspapers as counter-discourse, for example in the following excerpt from The Guardian:

(5) to put the figure in perspective, the total cost of benefit fraud last year was just £1bn. Corporate scrounging costs 11 times that

Source

1 that the Conservatives could restrict child benefit to just two children, to save around £2billion DMa

2 tax disability benefits or restrict child benefit to the first two children but the people G

3 obvious options, such as restricting child benefit to two children, cutting means-tested support G

4 contemplated the option of restricting benefits to two children, and never would. But why DT

5 cut child tax credit and restrict child benefits to children? Don't think Jenny was from DT

6 IFS said Labour's pledge to restrict child benefit for two years was "bizarre and indeed misleading DT

Concordance 7: Occurrences of benefit together with child* and restrict* in the Poverty-and-Benefits corpus

Source

1 insist that he did not want to cut child benefit or child tax credits as part of Tory plans DMa

2 said proved the Tories want to cut child benefits . Cameron said he rejected the plan at the DMa

3 believed a Conservative vow not to cut child benefit and tax credits. Just 16% said the Tories DMi

4 that the Conservatives want to cut child benefit is a major Labour attack line and the subject G

5 Cameron denies a secret plot to cut child benefit or tax credits Ed Miliband said last night I

Concordance 8: Occurrences of benefit together with child* and cut* in the Poverty-and-Benefits corpus

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

22 Journal of Corpora and Discourse Studies 1(1)

Similarly, the pattern of connection between RECEIVE and welfare fraud found partic- ularly (but not exclusively) in right-leaning tabloids in the Benefits corpus, was weaker in the poverty-and-Benefits corpus (Concordance 9), with a prevalence of articles focusing on the insufficiency and inadequacy of welfare (particularly in The Guardian — lines 3, 4 and 5), or on its fairness — for both taxpayers and claimants — (particularly in The Times — line 7). The Daily Mail (lines 1 and 2) presented welfare as opposed to work, but, this time, as a force that encourages (rather than discourages) people from work, as suggested by the title of the article (emphasis in the original):

(6) Generous welfare benefits make people MORE likely to work

Arguably, the capital letters in the headline suggest that normally welfare makes people less likely to work and betrays the same sense of scepticism expressed by the tabloid in the Benefits corpus. Thus, this article may be read as a way for the Daily Mail to dilute its usual narrative on welfare, but, it chose to present this inverted trend in an article where benefit* was mentioned in the vicinity of poverty.

Overall, this corpus showed a less critical portrayal of benefit claiming, with less em - phasis on benefit fraud, and on the welfare system as one that is ‘too generous’ and en- courages misconduct and dependency. The discourse surrounding poverty was more con- crete with more emphasis on in-work poverty and causes of poverty, such as unemploy- ment or low wages. Although marginal, this different representation suggests that the ideological distinction between deserving and undeserving poor, preserved in separate discussions of benefits and poverty, was disrupted when the two were discussed together.

Source

1 critical mass' of individuals receive public benefits rather than engaging in paid work, the DMa

2 may be because people who receive generous benefits when out of work may feel more inclined DMa

3 jobcentre was introducing to receive subsistence benefits , and persistent cold and hunger. Children G

4 reliant on welfare, some barely receiving benefits at all – about how five years of austerity G

5 the family no longer receives full child benefit – a loss that, somewhat to their surprise G

6 after loved ones; and taxing disability benefits received by more than 3 million people. G

7 perception that some of the out-of-work were receiving benefits which were excessive when compared with the earnings T

Concordance 9: Occurrences of benefit together with receiv* in the Poverty-and-Benefits corpus

7. Discussion The representation of benefit claiming in the press during the 2015 election campaign fo - cused on financial and domestic matters (such as proposed reductions of child benefits) and reflected the centrality of these issues in the discussion of welfare in the run-up to the

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

23

election. In contrast, when the discussion turned to poverty, there was little emphasis on public expenditure. While benefits were portrayed as something that can (and should) be dealt with at government level, poverty was framed in a global perspective, outside the realm of domestic issues. Confirming the findings of previous research (such as McKendrick et al., 2008), poverty was under-represented in comparison to benefit claims. Poverty was also presented as an intangible enemy endowed with agency and the ability to actively harm or trap people. It could be argued that this rendered any mention of the policies and structures that produce and sustain poverty superfluous. Similarly, by presenting the need to FIGHT poverty and HELP poor people as a moral issue, rather than a political or financial one, without, at the same time, discussing those who should FIGHT or HELP, the press offered both an emotional and vague portrayal that not only removed the blame from poor people for ‘being poor’, but also obfuscated the responsibility of those who should act to find solutions and end poverty. Thus, fighting poverty became a collective duty, in the absence of a specific entity responsible for ending it.

On the other hand, benefit* was talked about in terms of individual responsibility, rather than, for example, economic and wealth inequality. Unlike ESCAPE, the only verb co-occurring with poverty that implied the (partial) agency of poor people, benefit* co-occurred with LIVE OFF and GET OFF, which implied a greater sense of individual agency, personal responsibility, or even blame of the benefit claimant. The dichotomy between the representation of benefits and poverty recalled the deserving vs. undeserving poor ideology and the fact that escaping benefits — arguably more empathetic than getting off benefits — would be perceived as a marked expression in English shows how this contrast is reflected in and reinforced through language. However, as suggested by Spicker (2002, p. 7), representing claimants as ‘deviant’ individuals who ‘re - fuse’ to work and are ‘dependant’ on welfare, rather than as people who experience finan - cial hardship, is part and parcel of framing the welfare system as an impossible burden for the state and taxpayers and, therefore, in need of change.

This ideology is encapsulated in the phrase ‘getting off welfare and back to work’, an expression used by David Cameron at the 2014 Conservative party conference (British Political Speech, 2019), which presents work as the antidote to benefit claiming. The structure ‘getting off X and into Y’, constructs X and Y as lexical triggers of opposition (Jeffries, 2010, p. 48), thus constructing benefits and work as two parts of an imagin- ary whole in which the negative opposite is benefit claiming and the, undisputed, positive is being in work instead of, for example, tackling structural inequality. The opposition between those on benefits and those in work, like getting off benefits, placed the blame for relying on welfare onto claimants; this was in stark contrast with the repres - entation of poverty, where there was no suggestion that escaping poverty meant get- ting back to work. Additionally, this opposition distracts from the fact that many claimants need help, and do not simply ‘abuse the system’, taking ‘everything they can get,

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

24 Journal of Corpora and Discourse Studies 1(1)

often fraudulently’. It also reinforces, legitimises, and normalises the discourse of welfare dependency, its framing as a crime, and, consequently, the sense of urgent need for meas - ures (such as a welfare reform) to tackle fraud and the dishonest activities of those who claim benefits in order to ‘avoid’ work.

The dominance in the British press (and, consequently, in the corpora) of right-lean - ing newspapers (particularly tabloids), traditionally critical of welfare provision, can partly explain the prevalence of these themes in the Benefits corpus, given that right- leaning papers were over-represented in the Benefits corpus, both in terms of number of articles and words. Right-leaning tabloids alone represented the largest proportion of art- icles in the Benefits corpus (50.6%) and had the strongest focus on benefit fraud, with sig- nificant ideological implications in terms of prevailing discourse.

However, the analysis showed that traditionally right-wing ideologies (for example, aversion towards the welfare state and benefit claimants) found space in both left and right-leaning discourses and there was not enough evidence that, had the corpora been quantitatively balanced in terms of political allegiance, the results would have been qual - itatively different when poverty and benefits were discussed separately. The analysis of the major trends in the three corpora showed that a strong counter-discourse to the theme of benefit fraud and a dominant narrative of benefits as a positive way to support people were absent in the left-leaning press. Similarly, the left-leaning press did not focus on the inadequateness of the welfare system (Spicker, 2002) or on the discussion of in- work poverty, poverty wages and poverty pay. To the contrary, left-leaning newspapers tended to discuss poverty in absolute terms and as a geographically distant issue.

It was only in the poverty-and-Benefits corpus that the focus on scare stories of people abusing the system was backgrounded in favour of challenging the notion of wel- fare dependency and the assumption that large sums are regularly lost to benefits fraud. It appeared that a joint discussion of benefit* and poverty led both in the left-leaning and right-leaning press to a less accentuated emphasis on the ideological dichotomy between deserving and undeserving poor, thus leaving room for a representation that fo- cused less on a neo-liberal discourse of individual responsibility and more on the societal structures that cause poverty. However, this representation was peripheral and embedded within the dominant discourse of welfare dependency. It appears that separating the dis - cussion of poverty and benefit claims into two parallel debates allowed the press, for the most part, to rationalise and sustain two distinct arguments — benefits should be reduced (through welfare reform), but poverty should be ended.

Yet, seeing this separation simply as the direct result of the agenda of specific newspa - pers of different political leanings would be simplistic. The over-representation of right- wing papers alone does not explain the reduced focus on poverty, the distancing of the discussion of benefit claiming from discussion of poverty, and the different discourses surrounding poverty and benefits. Instead, these were more likely to be a reflection of the

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

25

fact that the leaders of the major parties on both sides of the political spectrum barely dis - cussed poverty explicitly and placed it firmly outside the arena of the campaign, as an is- sue detached from welfare, without tangible public impacts (Sippitt and Tranchese, 2015). Therefore, these representations can be seen as a symbiosis between political dis- cussions and journalistic practice, with the press regularly relying on the voices of institu - tional sources (Richardson, 2007), thus foregrounding their ideological stances and set- ting the agenda for the discussion of poverty.

This paper has demonstrated how, during the 2015 general election campaign, discus- sions of reductions to the welfare system as a cause of social inequality or as an exacerbat - ing cause of poverty (Spyker, 2002) were weak in the press, while discussions of welfare as a burden were dominant. It has argued that this reflected both the discourse of (oppos - ing) political parties and the lack of a strong left-wing voice in the British press, as shown by the intersection between media coverage and political discussions. At a time of major political change, like a general election, the press appeared to predominantly reproduce mainstream political ideology; in doing so, it also revealed a lack of a significant left-wing counter-discourse in the public debate on welfare.

Acknowledgements The author would like to thank Amy Sippitt for her work on the project.

Funding This paper comes out of research done for a Full Fact project funded by the Joseph Rowntree Foundation, to which the author was a contributor. Full Fact has not been in - volved in the preparation of this paper and all views are those of the author only.

Competing interests The author has no competing interests to declare.

References Baker, P., Gabrielatos, C., & McEnery, T. (2013). Discourse analysis and media attitudes: The

representation of Islam in the British press. Cambridge: Cambridge University Press.

Baker, P., & McEnery, T. (2015). Who benefits when discourse gets democratised? Analysing a Twitter corpus around the British Benefits Street debate. In P. Baker & T. McEnery (Eds.), Corpora and discourse studies: Integrating discourse and corpora (pp. 244– 265). London: Palgrave Macmillan.

Baumberg, B., Bell, K., & Gaffney, D. (2012). Benefits stigma in Britain. London: Elizabeth Finn Care.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

26 Journal of Corpora and Discourse Studies 1(1)

Bhatia, V. K. (1992). Pragmatics of the use of nominals in academic and professional genres. Pragmatics and Language Learning, 3, 217–230.

BNCweb (CQP-Edition). (2018). Retrieved from http://bncweb.lancs.ac.uk/.

British Political Speech. (2019). Leader's speech, Birmingham 2014: David Cameron (Conservative). Retrieved from http://www.britishpoliticalspeech.org/speech- archive.htm?speech=356.

Cabinet Office. (2010). State of the nation report: Poverty, worklessness and welfare dependency in the UK. London: Cabinet Office. Retrieved from http://www.bristol.ac.uk/poverty/downloads/keyofficialdocuments/CONDEM%20- poverty-report.pdf.

Charteris-Black, J. (2006). Britain as a container: Immigration metaphors in the 2005 election campaign. Discourse & Society, 17(6), 563–582.

Chauhan, A., & Foster, J. (2014). Representations of poverty in British newspapers. A case of “othering” the threat? Journal of Community and Applied Social Psychology, 24(5), 390–405.

Conservative Party. (2015). David Cameron speech to launch the Welsh manifesto. Retrieved from https://press.conservatives.com/post/116643787260/david-cameron-speech- to-launch-the-welsh-manifesto.

Department for Work and Pensions. (2011). The role of incapacity benefit reassessment in helping claimants into employment. London, UK: The Stationery Office Limited.

Dow Jones. (2019). Factiva: Global news database & licensed content. Retrieved from https:// www.dowjones.com/products/factiva/.

Fairclough, N. (1989). Language and power. London: Longman.

Fairclough, N. (2000). New Labour, new language. London: Routledge.

Ferragina, E., Tomlinson, M., & Walker, R. (2013). Poverty, participation and choice: the legacy of Peter Townsend. York: Joseph Rowntree Foundation. Retrieved from http://www.jrf.org.uk/sites/files/jrf/society-poverty-participation-full.pdf.

Freud, D. (2007). Reducing dependency, increasing opportunity: options for the future of welfare to work. An independent report to the Department for Work and Pensions. Leeds: Corporate Document Services. Retrieved from http://www.employabilityinscotland.com/media/82460/reducing- dependencyincreasing-opportunity-options-for-the-future-of-welfare-to-work.pdf

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

27

Government Digital Service. (2019). Benefits. Retrieved from https://www.gov.uk/browse/benefits.

Gries, S. (2008). Dispersions and adjusted frequencies in corpora. International Journal of Corpus Linguistics, 13(4), 403–437.

Halliday, M. A. K. (1985). An introduction to functional grammar. London: Edward Arnold.

Hart, C. (2014). Discourse, grammar and ideology. London: Bloomsbury.

Hoffmann, S., Evert, S., Smith, N., Lee, D., & Berglund-Prytz, Y. (2008). Corpus linguistics with BNCweb: a practical guide. Frankfurt am Main: Peter Lang.

Jeffries, L. (2010). Critical stylistics. The power of English. Basingstoke: Palgrave Macmillan.

Kelly, D. M. (1998). Stigma stories. Four discourses about teen mothers, welfare, and poverty. Youth Society, 27(4), 421–449.

Kilgarriff, A. (2009). Simple maths for keywords. In M. Mahlberg, V. González-Diáz, & C. Smith (Eds.), Proceedings of Corpus Linguistics Conference CL2009. Liverpool: University of Liverpool.

Kilgarriff, A., Baisa, V., Bušta, J., Jakubíček, M., Kovář, V., Michelfeit, J., Rychlý, P., & Suchomel, V. (2014). The Sketch Engine: ten years on. Lexicography, 1(1), 7–36.

Kress, G. (1994). Text and grammar as explanation. In U. Meinhof & K. Richardson (Eds.), Text, Discourse and context: Representations of poverty in Britain (pp. 24–46). London: Longman.

LabourList. (2014). “We will control immigration with fair rules” — Miliband announces Labour’s second election pledge. Retrieved from https://labourlist.org/2014/12/we-will- control-immigration-with-fair-rules-miliband-announces-labours-second-election- pledge/.

Leveson, L. J. (2012). An inquiry into the culture, practices and ethics of the press (report no. HC 780-II). London: The Stationery Office Limited.

Lexical Computing. (2019). Sketch Engine: language corpus management and query system. Retrieved from https://www.sketchengine.eu/.

Louw, B. (1993). Irony in the text or insincerity in the writer? The diagnostic potential of semantic prosodies. In M. Baker, G. Francis & E. Tognini-Bonelli (Eds.), Text and technology: In honour of John Sinclair (pp. 157–176). Amsterdam: John Benjamins.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

28 Journal of Corpora and Discourse Studies 1(1)

Lundström, R. (2013). Framing fraud: Discourse on benefit cheating in Sweden and the UK. European Journal of Communication, 28(6), 630–645.

MacDonald, R., Shildrick, T., & Furlong, A. (2014). In search of “intergenerational cultures of worklessness”: Hunting the Yeti and shooting zombies. Critical Social Policy, 34(2), 199–220.

McKendrick, J. H., Sinclair, S., Anthea, I., O’Donnell, H., Scott, G., & Dobbie, L. (2008). The media, poverty and public opinion in the UK. York: Joseph Rowntree Foundation. Retrieved from https://www.jrf.org.uk/report/media-poverty-and-public-opinion- uk.

Miller, G. A. (1995). WordNet: a lexical database for English. Communications of the ACM, 38(11), 39–41.

Mooney, G., & Neal, S. (2010). “Welfare worries”: Mapping the directions of welfare futures in the contemporary UK. Research, Policy and Planning, 27(3), 141–150.

Murray, C. (1990). The emerging British underclass. London: IEA Health and Welfare Unit.

Paterson, L., Coffey-Glover, L., & Peplow, D. (2016). Negotiating stance within discourses of class: reactions to Benefits Street. Discourse & Society, 27(2), 195–214.

Princeton University. (2010). Wordnet: a lexical database for English. Retrieved from https://wordnet.princeton.edu/.

Pykett, J. (2014). Representing attitudes to welfare dependency: Relational geographies of welfare. Sociological Research Online, 19(3), 23.

Rayson, P. (2002). Matrix: A statistical method and software tool for linguistic analysis through corpus comparison. Unpublished doctoral dissertation, Lancaster University, Lancaster.

Richardson, J. (2007). Analysing newspapers. An approach from Critical Discourse Analysis. New York: Palgrave Macmillan.

Savický, P., & Hlaváčová, J. (2002). Measures of word commonness. Journal of Quantitative Linguistics, 9(3), 215–231.

Sippitt, A., & Tranchese, A. (2015). How did the media and politicians in the UK discuss poverty in 2015? York: Joseph Rowntree Foundation. Retrieved from https://www.jrf.org.uk/report/how-did-media-and-politicians-uk-discuss-poverty- 2015.

Smith, K. (Producer). (2014). Benefits Street [Television series]. London: Channel 4.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9

29

Spicker, P. (2002). Poverty and the welfare state: Dispelling the myths. London: Catalyst.

Spicker, P. (2007). Definitions of poverty: Twelve clusters of meaning. In P. Spicker, L. Álvarez, & G. David (Eds.), Poverty: An international glossary (pp. 229–243). London: Zed.

Stubbs, M. (2001). Words and phrases: Corpus studies of lexical semantics. Oxford: Blackwell.

Townsend, P. (1979). Poverty in the United Kingdom. Harmondsworth: Penguin.

Welfare Reform and Work Act (2016). Retrieved from http://www.legislation.gov.uk/ukpga/2016/7/section/7/enacted.

Tranchese (2019) Getting off benefits or escaping poverty? DOI 10.18573/jcads.9