The Role of Discourse Analysis in Researching Severe Labour Exploitation

Investigating severe forms of labour exploitation presents a series of particular methodological challenges to researchers in the field, including access to respondents, credibility of data, reliability of measures, researcher ethics and the practical and political dimensions of study design. For researchers embarking on qualitative approaches – whether it involves interviews, ethnography and/or documentary forms of analysis – this chapter seeks to illuminate the potential of a discursive approach to understanding severe forms of labour exploitation. It aims specifically to help understand how severe forms of labour exploitation are variously constructed as an object of knowledge/s, and how this construction is always contingent upon socio-political con/texts. To this end it recommends the investigation of texts as data, proceeding to discuss some interpretive work generated from an early-stage analysis of media, government and civil society discourses surrounding the UK Modern Slavery Act 2015.


Researching modern slavery
The earlier chapters in this book attest to the challenges presented to those who take severe labour exploitation as a central problem in their research. Each challenge is an outcome of the particular research trajectory that is being adopted, the nature of questions being asked, the unit of analysis pursued, and the political dynamics involved in the production of research. Efforts to provide a clear definition of forced labour, for instance, are met by the historical 'slipperiness' of the language used to describe it (Dozema 2000), what principal actors such definitions include, and the kinds of practices that it refers to (Quirk 2011). Similarly, attempts to measure incidences of modern slavery encounter problems of data reliability, and credibility of claims made about its scale, distribution and frequency. Questions concerning reliability also extend into research that has prioritized the victim as the central unit of analysis. Whilst it may by possible that 'getting closer to the victim' yields better research, the continued focus on a victim perspective has marginalized the value of research into the perpetrators of extreme labour exploitation (Crane 2013). Finally, as several Chapters argue, the political conditionality of data production, interpretation and implementation, such that what comes to pass as the 'truth' about modern slavery may be influenced by organisational interests.
In light of these varied methodological challenges, the present chapter outlines the value of adopting discursive methodological approaches to researching severe labour exploitation, as one strategy to overcome some of the persistent challenges just presented. Whilst it does not claim superiority over other methods or immunity from the aforementioned challenges, it does open up an alternative research trajectory through which interested scholars might apprehend severe labour exploitation as a distinct object of research.
Discourse analysis concerns itself with the way in which language is deployed by social actors to produce knowledge about particular phenomena (e.g. modern slavery). In this sense, when thinking about definitions for instance, discourse analysts would be interested in how certain definitions come about, why some become accepted, how others are rejected and/or marginalised, as well as how definitions shape the subjects and practices we come to associate with severe labour exploitation and how this is influenced by socio-political practices. Similarly, because discourse analysis is concerned with investigating 'texts' rather than subjective experiences or verbal accounts, it doesn't face the same challenges of data access, reliability, and researcher ethics. However, this does not mean discourse analysis is an easy option into a difficult phenomenon. Rather, exploring written accounts, visual representations, annual reports, social media, policy documents, websites, poems, songs, and even photographs can, in certain contexts, be more methodologically practicable, appropriate, and revelatory than talking to the victims themselves.
In this chapter, I will be arguing that discourse analysis helps address the question of how severe labour exploitation is produced as an object of knowledge (Fairclough 2013), and how this shapes the subjects and practices that come to be associated with it. In this socially constructivist view, notions such as 'modern slavery' are not taken as a natural, real or fixed phenomena, but as linguistic forms, whose meaning is contested by social actors with respective interests in the form they adopt. Modern slavery is, accordingly, seen here as part of a 'discursive struggle' (Livesey 2001) over what severe exploitation means, who it involves and what kinds of remedial practices this articulates for actors in public discourse (Dahan & Gittens 2010). Drawing from a concise cluster of methodological literature, I will outline some of the key conceptual properties of discourse analysis that may be fruitfully mobilized. These will then be applied by drawing upon some textual excerpts from the UK Modern Slavery Act 2015, to demonstrate how different actors with vested interests (Corporations, Media and NGOs) attempt to construct 'Modern slavery'.
Through this, severe labour exploitation can be understood as a heterogeneous concept constructed around distinct sets of subjects and practices that are contingent upon actor interests. The following section will consider in more detail what discourse is, and how it might be analysed accordingly.

What is discourse analysis?
To provide a concise, capture-all definition of discourse analysis would be to go against the consensus that it constitutes a fairly wide variety of approaches, theoretical assumptions and methodological techniques (Alvesson & Karreman 2000). Moreover, it can be used in conjunction with a number of other qualitative methods such as ethnography (Covaleski et al. 1998), narrative analysis, content analysis, photographic analysis, conversation analysis (Wetherell & Potter 1992), interviews, and other archival documentary forms of analysis. This diversity of methods is matched by the potential scope of phenomena that might become the subject of a discourse analysis.
Political speech, gender, race, ethnicity, poverty, social control and exclusion, organisational change, immigration, and a wide variety of organisational practices have all been subject to some form of discourse analysis. Given this variety of methods and phenomena, what are the properties that actually differentiate discourse analysis from other qualitative techniques?
Discourse analysis concerns itself with the investigation of processes of social construction (Berger & Luckmann 1966). Analysts explore the way in which knowledge of social reality is produced through the purposive use of language (Fairclough 1995). Discourse -as the purposive use of language-is not seen as a benign reflection of a hidden, 'real' social mechanism that determines the behaviour of social actors. It is inherently active in shaping the knowledge of those social actors in the first instance. Discourse analysts cannot therefore assume phenomenon such as 'modern slavery' or 'forced labour' are a priori categories, reflective of underlying naturally-occurring social mechanisms. It is more the case that notions such as 'modern slavery', 'chattel slavery' and/or 'white slavery' constitute specific linguistic representations that have distinct social functions and historical contexts: Discourse analysis is concerned not with specifying what sentences are possible or 'grammatical', but with specifying sociohistorically variable discursive formations (sometimes referred to as 'discourses'), systems of rules which make it possible for certain statements but not others to occur at particular times, places and institutional locations. Fairclough (1992: 40) This is not to say that severe forms of labour exploitation do not exist as a subjective experience of a victim, for example, but the way in which it becomes an object of knowledge -about certain kinds of subjects (victim, perpetrator) entailing specific kinds of practices (violence, deception, incarceration) is a product of discourse.
In this chapter I will work from the basic idea that discourse constructs knowledge ('ways of knowing') about social identities or subjectivities ('ways of being') and social practices ('ways of doing') (Caruana & Crane 2008;Parker 1992).
To illustrate what I mean by this, we can consider in turn Foucault's treatise of the social categories of 'madness' and 'crime' with Joel Quirk's historical account of the anti-slavery movement. Foucault (1972) argues that the category of 'mental illness' was socially constituted in the contested discourses of psychopathology, through all the 'statements that named it, divided it up, described it, explained it…(1972: 32).' Crucially, Foucault suggests that discourse produces knowledge of subjects and practices -not only connecting who subjects are with what they can do (Parker 1992) but instructing what practices can be done to them. Thus, where the subject 'accused person' moves through the justice system into the position of 'guilty criminal', their 'body' becomes subjected to institutionalized penal practices of incarceration, punishment, isolation and rehabilitation. We can approach notions such as 'modern slavery' in a similar manner -as a category of knowledge about severe labour exploitation involving certain kinds of distinct subjects and practices. For example, whereas for subjects of human trafficking, their exploitation may be deeply connected to practices of movement, social alienation, and isolation, for debt bonded labour, such practices are, conversely, associated with the lack of physical movement and the presence of deeply embedded class structures.
It is crucial to note that such subjects and practices do not just appear in or spring from discourse. Their form has context/s; emerging gradually through processes of contestation between powerful social actors. We see this in Quirk's (2011) account of the anti-slavery project in which, following the legal abolition of slavery, (now) 'free labour' was often subjected to novel coercive practices of exslave owners and politicians: Throughout the nineteenth and early twentieth century, both former slaver owners and policy-makers regularly turned to other forms of coercion and compulsion to secure labor on their preferred terms…many strategies were developed in pursuit of these goals, including forced labor schemes, taxation in the forms of labor, draconian labor 'contracts', forced recontracting, indentured labor schemes, debt bondage (and) restrictions on movement and land ownership…. Quirk (2011: 121) Echoing Foucault indirectly, Quirk's historical account highlights the point that severe labour exploitation has been subjected to socio-political contestation that is responsible for shaping, again, what subjects fall into 'slavery (like)' categories and, in turn, what practices can be done by and to them. The value that discourse analysis brings to this complexity, therefore, is in the investigation of processes of social construction in which forms of labour exploitation are constructed and legitimated (Wetherell & Potter 1992) and how these shape organisational and institutional responses to them. I elaborate further on this political view of discourse as the chapter unfolds and demonstrate this in relation to my research on the Modern Slavery Act 2015. In the next section we turn our attention to the mobilisation of discourse analysis.

Doing discourse analysis
I have previously stated that there is no one-size-fits-all approach to the analysis of discourse and that there is a manifest variation in approaches that reflect the nature of the research agenda, ontological assumptions of the researcher, and not least the object of analysis itself. In deciding upon which 'variety' of discourse (Alvesson & Karreman 2000) we might apply, it is helpful to first understand the relationship between a researcher's theoretical trajectory and the distinct properties of discourse. Phillips and Hardy (2002) helpfully summarize the varieties of discourse analysis available along two continuums. The first continuum reflects the degree to which researchers are concerned with processes of social construction -ie. how realities are assembled through discourse -as opposed to power and ideology -i.e. those powerful actors that shape and constrain that construction of reality. Thus we might discern between studies that focus on how slavery is communicated to consumers -via websites, adverts and product labels -to help them understand what 'fair trade' actually means (McDonaugh 2002) and those studies that observe how contemporary forms of slavery are contested through discursive struggles (Livesey 2001) between NGOs and corporations, for example (Dahan & Gittens 2011). The second continuum maps the degree to which researchers are interested in the discursive dynamics within a micro 'text' -such as a single social report, advertisement or corporate website -as opposed to concerns with the broader, macro social context from which such texts draw their meaning. 1 It is important to note that these are continuums and are not dichotomous in the sense that researchers focused on the construction processes in texts might also have some interest in their broader socio-political context. In this vein, Fairclough (1995) has argued extensively for a three-dimensional approach to discourse analysis in which the micro linguistics of a text should be analysed in relation to the meso and macro level socio-political practices in which they are simultaneously embedded. This is evidenced neatly in Dahan and Gittens' (2011) study of public responses to child slavery in the cocoa industry. They show that a 'responsible business' frame for addressing labour exploitation is created by vested (cocoa) industry actors as a defensive response to a 'fair trade' frame produced by NGOs, that is critical of current business practices. The underlying tensions within these competing discursive frames can be summarized in terms of public calls for more extensive accountability and regulation of business and a countervailing corporate agenda of voluntary selfregulation. In this sense, how the public understands what slavery is/is not has direct implications for the subjects and practices of both NGOs and corporations in that it creates expectations for what constitute appropriate remedial actions (e.g. stricter regulation vs self-reporting). In exploring such 'discursive struggles' we are reminded again that no forms of severe labour exploitation are natural or objective phenomenaat least not to the discourse analyst -but emerge as contested categories of conditioned meaning as will be demonstrated further below.
Before proceeding to a deeper exploration of discourse analysis as a method, it is necessary to make explicit the status of text both conceptually -as containing tissues of cultural meaning -and methodologically -as the site of data. Firstly, discourse operates in and through texts. It is in and through texts that processes of social construction occur (Phillips & Hardy 2002). Texts are linguistic, social and ideological vessels, in which ways of knowing are assembled with a particular audience in mind (Fairclough 1992). Thus texts are key sites in which knowledge about modern slavery are produced, disseminated and consumed. Moreover, a broad view of discourse makes no qualitative distinction between the value of spoken, written of visual texts. This means that a debate in the House of Commons, a legislative act or report, or even a photograph can help discourse analysts understand how knowledge about severe labour exploitation is constructed. This is because of the property of 'inter-textuality' (Fairclough 1992), which simply means that texts are always fragments of meanings drawn from other texts, all of which channel and reify certain strands of discourse. So, for example, studying a single company's published corporate social responsibility (CSR) report before and after the UK Modern Slavery Act 2015 was passed, may reveal how knowledge of modern slavery has become institutionalized within business auditing practices and whether this has shaped the subjects and practices that businesses associate with severe labour exploitation.
This conceptualization of discourse analysis opens up a whole range of texts that might be usefully investigated as sites of data collection. Contingent upon the object of research and nature of research questions, corporate reports, NGO and activist websites, photographs, journalistic articles, political speeches, public debates, poetry, songs and even art significantly extend (and to my mind compliment) existing research favouring victims' verbal accounts of severe labour exploitation. Moreover, by opening up the boundaries around the data in this way, it is possible to explore the socio-political context/s that shape and orientate knowledge of severe labour exploitation in particular ways. This will be analysed in more detail in the remainder of the chapter.

Constructing Severe Labour Exploitation
In this next section I want to emphasize the situated nature of texts in understanding the construction of knowledge about severe labour exploitation. That is, the idea, that what is being articulated in discourse on modern slavery is a product of who is talking/writing (text producer) and the intended audience (text consumer), coupled with the specific agent interests that may shape this process. In structuring each section, I first point to key insights from other, thematically similar discourse-based studies to elaborate the theoretical lens being used, before moving on to synthesise this in the context of severe labour exploitation. For the latter, I will draw upon some of raw data from an ongoing discourse analysis of the UK Modern Slavery Act 2015 that I am undertaking with Andrew Crane and Claire Ingram. Specifically I will draw out some of the discursive forms -particularly subjects and practices -that the media, government and NGOs used to articulate severe forms of labour exploitation, whilst contextualizing the form of discourse in each context.

Constructing Subjects: Media texts
The construction of subjects associated with modern slavery (e.g. slaves, victims, perpetrators, etc.) is contingent upon who is talking (or writing) about them.
Moreover, those organisations constructing subjects of modern slavery (e.g. who is a 'typical victim' or 'perpetrator') tend to do so by their access to pre-existing discourses (e.g. 'sex workers' or 'criminal gangsters'). The property of 'intertextuality' (Fairclough 1995) is important here, in that subjects like modern slaves, migrants or refugees do not simply emerge from nowhere, but are rewoven into the sematic fabric of new texts, rendering novel configurations of subject-types possible. For example, Phillips and Hardy (1997) argue that the 'refugee' subject is constructed as a subject of knowledge from other discourses: In the case of the UK Refugee system, discourses that produce a refugee [as a subject] draw on other discourses (the UN's Universal Declaration of Human Rights, for example) and includes a wide range of texts including government reports and statements, news reports, cartoons, editorials and demonstrations. Phillips and Hardy, (1997: 166) I illustrate here how the subject 'modern slave' is, similarly, constructed from other discourses and that these constructions reflect the local con/text, in this case, the socio-political leanings of the UK mainstream press. In this section I will demonstrate how different mainstream media in the UK attempted to construct the 'modern slave', as a distinct subject. In each case, it is evident that the linguistic form of each text anticipates interpretations from a specific audience, such that the common interests of The Daily Mail readership give form to the textuality of modern slavery in a way that is distinct from The Guardian readership. Whilst each media actor builds a picture of modern slavery of (attention-grabbing) interest to their audience, we can also note a tension playing out across all actors between the degree of exploitation that subjects of modern slavery face.
The passage below from The Independent discursively builds a picture of the victims that are normally subjects of modern slavery. The focus of the article is upon the disproportionately high number of women and children (as opposed to men), underlying the vulnerability of subjects. Moreover, the passage connects these vulnerable subjects to very extreme, criminal practices involving sexual abuse and human trafficking.
More than 140 children were identified as being trafficked for sexual abuse last year as the numbers of people rescued from being held as slaves soared by nearly half. The child victims included 88 youngsters brought in to Britain from overseas, most commonly from Vietnam and Albania, and another 56 who were UK-born….. Sixty-four per cent were female and 36 per cent were male, while 26 per cent were aged under 18. The increase emerged as the Home Office prepares for a major crack-down on modern-day slavery. Theresa May, the Home Secretary, has said she wants tough legislation in place to combat it by the time of the next election in May 2015. The NCA disclosed that half of the women smuggled into the country worked as prostitutes, with smaller numbers forced into domestic servitude or laboring……in the sex industry (emphasis added). [The Independent, 18th February 2014] This passage, like others in The Independent articulate modern slavery as involving highly vulnerable subjects, facing extreme (mainly sexual) exploitation.
In the next sample from The Daily Mail we see other known-to-the-readership, criminal subjects being associated with modern slavery. Whilst the passage below echoes the metaphorical template of the previous passage in The Independent -that modern slaves are normally victims of sexual abuse -it extends its account to elaborate more on the subjectivity of the perpetrator: The indictment alleges that women and girls were recruited from city streets or social media to join the sex-trafficking ring. The victims then had to deliver their earnings to pimps in exchange for protection, food, housing, clothing and cars. The network was run by a gang known as BMS, which traces its origins to San Diego's increasingly gentrified North Park neighborhood in the early 1990s. BMS members have nicknames like 'Pimpsy' 'Stick Up' and 'Li'l Play Doh,' prosecutors said…..(emphasis added).

[The Daily Mail, 9th January 2014]
Here, and in other media texts (below), we start to see the invocation of organised crime in the guise of gangsters, pimps, and the like. This acts as a discursive template for the audience, helping them to understand the kinds of perpetrators involved in severe labour exploitation.
When we look closer at the text, beyond gangster metaphors, we can see other linguistic devices such as normativity and moral panic. In the passage below from The Telegraph, the audience is initially panicked by the potential scale of the problem -'tip of the iceberg': The discovery of three women allegedly held as 'slaves' for 30 years is the "tip of a rather large iceberg", according to an MP in charge of reviewing evidence of slavery in Britain. Frank Field, chair of the Modern Slavery Bill evidence review, said criminal gangs were making "huge sums of money" from people being imported into the UK to work "almost for nothing". "We've had this example of domestic slavery but people are being imported to work, almost for nothing, in industry," he said. "We've got begging gangs being developed, with people being imported. And of course we've got the whole question of how children are being imported to work.."….. "It's so shocking to just be talking to people who have been through this," the Labour MP for Birkenhead said……It is thought that the 30-year-old woman had been held captive all her life….(emphasis added).
[The Telegraph, 23rd November 2013] Rather than pointing to 'varied types' of modern slavery subjects, the use of the term 'tip of the iceberg' highlights a very extreme example ('life-long' incarceration), and suggests it is potentially prolific.
Finally, when we talk about how modern slavery is being defined, we can also see how these metaphorical templates (e.g. organised crime) invoke practices by other institutions (e.g. the government and/or corporations). In the example below, by referring to modern slavery as an organised crime committed by gangsters, this invites conventional remedial practices such as the seizing of assets from gangsters: Human!trafficking!gangs are getting away with millions of pounds in profits because the police have no incentive to freeze their assets, Home Secretary Theresa May is warned today. MPs warn new laws to tackle modern-day slavery do not enough to hit gangmasters 'where it hurts' by confiscating cash!immediately. They urge Mrs May to follow the example of US gangster Al Capone who was 'brought down in the 1930s through his finances. (emphasis added).

[The Daily Mail, 4th November 2014]
In all of these short examples, the subjects associated with modern slavery are extreme and highly criminalised. The Guardian offers something of a juxtaposition to this emerging view of modern slavery as involving extreme subjects only. They, in contrast, tend to articulate to their audience modern slavery as being on a continuum of exploitation, involving a more nuanced range of subjects from the outright extreme to more everyday: The committee notes that modern slavery in the UK ranges from the exploitation of adults and children in the sex industry to forced labour, domestic servitude and such forced criminal activities as cannabis farming. It says victims include British schoolchildren, children brought to the UK for benefit fraud and those who are trafficked or come to the country legitimately and voluntarily only to find themselves subsequently enslaved (emphasis added).
[The Guardian, 8th April 2014] The direct use of the term 'ranges' indicates a broadening view of the types of subjects that might be associated with modern slavery. There is apparent variation in the status of those who might find themselves as victims, with some travelling for legitimate and self-determining reasons (i.e. not trafficked as prostitutes).
In continuing the broadening theme, The Guardian associates modern slavery here with potentially 'legitimate' actors and spaces (i.e. not only organized crime), such as supply chains and international corporations: These reports show how businesses of all kinds are sullied by slavery -…The reports pierce several myths. Many believe that workers who knowingly migrate without papers cannot possibly be victims. But traffickers' force, fraud, and coercion can make them just that. Some people, conversely, assume that only undocumented migrants can be victims of human trafficking. But many legal guest workers are put so deeply into debt by recruiters, robbed of their autonomy and subjected to such harsh work that they become veritable slaves. Many people suppose that most trafficking is for sexual exploitation (as in Thailand's own sex trade)…. ….And many executives assume that tracing labour conditions in their supply chains is futile and prohibitively costly. Yet if the Guardian can do it, the business community surely can -and must (emphasis added). [The Guardian, 27th June 2014] In this extended passage, it is proffered that businesses of all kinds (not just drug or sex gangs) might be subjects potentially associated with modern slavery. Moreover, even legally recruited guest workers might be subjected to practices that are sufficiently similar (Quirk 2011) to be termed as 'veritable slaves'. The myth-busting narrative that they employ through the latter part of the passage acts as a debunking device, aimed at undermining narrower, extreme constructions of modern slavery we saw in the earlier examples. This is most likely because The Guardian is championing more extensive legislation and corporate accountability in relation to this issue, which is in alignment with is support for ethical trading initiatives, corporate responsibility and sustainability -key interests of its center-left audience. And this point is crucial to any discourse analysis of modern slavery. It is not just the subject (modern slave or refugee) that is being constructed in the text but the organizational identities of other agents implicated in them: In the case of refugee discourse, it is not just refugees that are produced; so, too, are the immigration officers that admit them; the decision-makers who determine their status; the members of NGOs who provide them with services; the media which report on them; the public who read about them. Phillips and Hardy (1997: 169) Indeed, modern slavery is not only about extremely vulnerable 'victim-subjects' who are shot, raped, or sold for sexual exploitation, and can be equally associated with other subjects such as low-paid workers, governments, media and legitimate corporations such as the readership's much favoured John Lewis: If John Lewis, with its exemplary anti-slavery processes, can find exploitation in its supply chain, what hope for this modern slavery bill?
Last year, 30 miles up the road from my constituency, more than 40 Hungarians were found working for less than £2 a day in a mattress factory in Dewsbury, and living in squalid conditions. Crammed into a two-bed flat, they were surviving on food scraps, and were threatened with violence if they complained (emphasis added). [The Guardian, 17th November 2014] In the context of the previous articles, passages from The Guardian provide a countervailing force that seeks a wider range of subjects (both victim and perpetrator) and thus more extensive apparatus for government (legislative requirements for reporting) and corporations (supply-chain auditing) to enact in its remediation.

Institutionalizing practices: Government Texts
If the previous section highlighted the actor-contingent context for constructing modern slavery as a certain kind of subject, this next section elaborates on what it can be as a practice. However, rather than looking at practices connected to the specific subject 'modern slavery' (e.g. threats, deception, etc.) we move to a higher level abstraction and consider the role of discourse in institutionalising certain practices (e.g. auditing, reporting, legislating). In short, this section highlights how severe labour exploitation can become constructed as an object of particular institutional arrangements. In their exploration of slavery in the cocoa industry, Dahan and Gittens (2010: 227) view labour exploitation as an 'ethical public issue' that is socially constructed through framing contests between strategizing actors such as the government, firms, NGOs, and the media; each actor vying to impose their preferred solution to the issue. In this excerpt from their data analysis they view NGOs' promotion of mandatory fair trade certification scheme upon the whole cocoa industry as the best solution, as follows: This is why most involved NGOs have rather advocated an extensive ''fair trade'' certification as the best solution. This is exemplified in this letter addressed to Mars Masterfoods: ''Through the steps outlined in the Protocol, there is no guarantee that prices will rise to sufficient levels and remain stable. Fair Trade, in comparison to the projects you are funding, is truly a 'much more holistic approach' (emphasis added). Dahan and Gittens (2010: 236) In direct response, the industry -resistant to external governance -contests the viability of mandatory certification: 'Fair trade' is one of several options to consider. Ultimately, however, Fair Trade is an approach that works best with farms that have access to infrastructure such as communications and warehousing facilities. While our long-term goals include encouraging the development of farmer organizations, currently the majority of farmers in West Africa do not have access to the type of infrastructure that is needed to take part in a fair trade supply chain. In this first passage below, we see a clear orientation to the extreme view of modern slavery that is being circulated in some of the more centre-right media (e.g. as serious, organised crime). The passage is very explicit that only the most extreme subjects are to be included in the identification of modern slavery: identifying a clear 'tipping point' at which it becomes an unequivocal criminal offence: I appreciate that, and that is exactly where we want to be at the end of this process: that there is a simple-to-understand offence for law enforcement agencies and others, which has an appropriate sentence with it. Clearly, one of the concerns is that we don't want to end up with relatively minor offences being brought into the definition. This is an offence that carries a life sentence. This is a very serious offence. You talk about the beds-and-sheds-type example and also miracle babies. I would be very keen, if you have any suggestions, to make sure that we can capture those offences, because, clearly, if you asked a member of the general public what they mean by modern slavery, an offence of taking a baby without permission would probably come within that somewhere. I think we do want to make sure-I don't just think: I know-that we are covering those offences, while keeping the Bill focused and making sure it is the serious offences that we are tackling here (emphasis added).
[Karen Bradley MP, Minister for Modern Slavery and Organised Crime, Home Office, examined.] This extreme view of modern slavery continues in the passage below. However, I wish to emphasize more here, the practices in the form of institutional responses that are being promoted. In particular, the government is underlining the importance of non-intervention in labour markets; providing an implicit distinction between corporate governance via new or extended legislative over corporations as opposed to voluntary business practices (e.g. reporting): I know how important it is for businesses to play their part in tackling modern slavery. I am committed to working with business to eliminate modern slavery from supply chains. That is why we have included a world-leading disclosure requirement in the Modern Slavery Bill to require all large businesses to disclose what they have done to ensure their supply chains and own business are slavery free. This measure will harness consumer and investor pressure by giving them clear information about what action businesses are taking and will drive businesses to do more to ensure that they do not unknowingly encourage these heinous crimes." Many businesses are already taking action to eliminate modern slavery. However, a range of NGOs, Parliamentarians and businesses have also suggested that a specific disclosure requirement focused on modern slavery would be a nonburdensome way of increasing transparency further and encouraging businesses to take more action. The Government has been considering these representations carefully, to determine if improvements could be made without over-complicating existing arrangements (emphasis added). From this we could make the argument that the discourse is working to incorporate modern slavery into existing institutional practices of both corporations and governments, rather than changing them. Corporations, for example, are required only to include modern slavery into existing reporting practices whilst governments, for their part, are similarly attempting to accommodate modern slavery into legislative mechanisms that already exist for high profile crimes. In the forward to the Modern Slavery Act it is stated that 'The Bill will consolidate the existing slavery and human trafficking offences into one Act of Parliament.' In this sense we could start to assemble a more critical argument, if so inclined, for the discourse of the 'status quo' (Livesey 2001): the absorption, simplification, and normalisation of something novel into pre-existing, 'known' institutional practices.

Revealing 'realities': NGO texts
Because discourse constructs 'ways of knowing' (Foucault, 1972) about the kinds of subjects and practices that, in our case, constitute severe labour exploitation, studies can also consider how dominant constructions stand in for, mask over, other 'ways of knowing'. This notion connects with research that considers the powerful effects of 'representations' or 'myths' (Doezema, 2000) in obscuring other potential realities. In their study of New Zealand's tourism representations, Ateljevic and Doorne (2002: 648) regard "imagery as a political process that encodes and reinforces the dominant ideology of tourism culture", culminating in the subversion, systematic exploitation and exclusion of Maori communities. Their study draws on a range of historical marketing texts that the tourism board has used to market the country to an international tourism audience. Representations of New Zealand successfully depict a romanticized view of the island and its rich geographic and cultural resources, whilst at the same time presenting, for example, relationships between the (colonising) Anglo-Saxon population and indigenous Maori population as civilized and 'trouble free'. Such romanticized representations of the exotic, effectively, mask a historic reality: The reality of Maori as largely urbanized people suffering high levels of intergenerational unemployment, poverty, and incarceration rates are carefully avoided by the contemporary tourism discourse. Ateljevic and Doorne (2002: 662) From this we can consider that as a dominant understanding of what constitutes severe labour exploitation comes to settle in public discourse as the view, it may well (perhaps inadvertently) mask other realities. Our research suggests that this is anticipated by many of the NGOs who are cautious about the narrow, crimefocused constructions of modern slavery. As reported in the more left oriented media outlets, there was concern from civil society actors that the 2015 Act was rather too narrowly defined to be effective: Without the use of indicators of modern slavery, it is unclear how cases other than the most severe and obvious will be identified…..The measures in the draft Bill would be made more effective, with regard to both prevention and enforcement, by the incorporation of such indicators, either in the Bill itself or in accompanying guidance (emphasis added). [CORE Coalition 2014: 71] As distinct from the extreme view, where Modern Slavery is regularly depicted in the context of extreme illegitimate spaces (e.g. rape, torture, organized crime, the sex industry), NGOs often saw them potentially occurring within more legitimate ones, and involving a broad range of practices -not limited to baby theft or murder: In the UK food sector, research identified at least 14 forced labour practices, including: up-front fees/debt bondage, threats and bullying, disciplining through dismissal, overwork, non-/under-payment of wages, underwork/indebtedness, deductions/charges, documentation abuses, tie-ins (work permits, accommodation, money) (emphasis added).

[Joseph Rowntree Foundation]
Such passages attempt to articulate a broad scope of subjects and practices associated with modern slavery, pointing to the potentially greater scale, pervasiveness and ambiguity. This framing is far removed from the rather static, extreme view that recommends a clear and simple understanding of only the most serious offenses.
Consequently, by taking these broadening constructions of modern slavery, we are able to ask more critical questions about the ongoing tripartite 'discursive struggles' (Livesey, 2001) between corporations, government and civil society actors whose interests and goals shape how they attempt to construct the reality ('regime of truth') around severe labour exploitation, rendering it as a particular object of knowledge.

Concluding thoughts
As other chapters in this book attest, investigating severe forms of labour exploitation presents a series of particular methodological challenges to researchers in the field.
In this chapter I have tried to recommend an alternative trajectory for researchers pursuing more qualitative research design, that prioritizes the role of 'texts' in understanding severe labour exploitation. As discussed, this may require an ontological move on the part of the research who, rather than taking notions such as 'modern slavery' or 'forced labour' as a priori categories, instead seeks to understand the processes that socially construct the category; that bring it into being as a certain object of knowledge. Having opened the door to a discursive approach, a wide variety of discourse analysis techniques may be deployed that are more or less contingent upon the unit of analysis, object of research and, of course, the nature of the researcher's questions. Whilst these varieties of discourse analysis may convey certain methodological orientations -e.g. towards micro linguistic processes or macro-political power structures -these are not necessarily dichotomous tendencies.
In this chapter, I have suggested that researchers, if so inclined, may usefully combine different levels of analysis such that the linguistic functions of a text (e.g. metaphor, narrative, juxtapositions) can be interpreted in the context of interpretive effects on particular audiences, and even the macro socio-political forces that shape them. This was unfolded to an extent in considering the different configurations of subjects and practices that were used by different social actors when constructing modern slavery.
Ultimately, what this chapter has tried to do for those embarking on qualitative modes of inquiry, whether it involves interviews, ethnography and/or documentary forms of analysis, is to help them understand how severe forms of labour exploitation are variously constructed as an object of knowledge/s, and how this construction is often contingent upon a socio-political con/text.