Discourse and Migration

Since the 1960s the multidisciplinary study of discourse has become increasingly popular in the humanities, as well as the cognitive, social and political sciences. This new cross-discipline offers many sophisticated quantitative and especially qualitative methods for the explicit and systematic analysis of text and talk, including the analysis of syntax, semantics, pragmatics, narrative, argumentation, style, corpora, metaphor, ideology, and multimodal analysis of images, among others. These methods have also been applied the fields of the study of migration, ethnic relations, minorities and racism. In this methodological chapter, various methods of discourse analysis are illustrated with passages from a variety of discourses on minorities and migrants.


Introduction
This chapter introduces some discourse analytical methods for the study of migration. It should be stressed from the outset, though, that discourse analysis is not a method, but a broad, multidisciplinary field of study of the humanities and social sciences, a field that therefore should rather be called Discourse Studies. This field has emerged from developments, since the mid-1960s, initially in anthropology, (socio)linguistics, literature, semiotics, cognitive psychology and sociology, and later in social psychology, history, law and communication. Strangely, political science is one of the few disciplines in the social sciences that has not witnessed such a popular development, although many if not most political phenomena are discursive. Today, after more than half a century, Discourse Studies is a broad cross-discipline, with thousands of books and articles, half a dozen journals and conferences on many specialized topics (for general survey to the whole field, see Tannen, Hamilton & Schiffrin, 2015; for introduction see Van Dijk, 2011; for further references, see below).
In this field many methods are used, such as analyses of context, gestures, sounds, syntax, meaning, rhetoric, speech acts, interaction, conversation, narrative, argumentation or genres such as news reports or parliamentary debates, each again with more specific methods and approaches. In general, these methods study different kinds of structures and strategies of text and talk. Besides these more qualitative methods, there are more quantitative methods, such as corpus linguistic methods of vast text corpora, experimental methods in the psychology of discourse processing, or ethnographies for the study of discourse in its social and cultural contexts. Qualitative methods of discourse analyses are different from traditional Content Analysis, which is generally based on quantitative treatment of coding of discourse expressions, using ad hoc codes, that is, without an explicit theory of discourse structure (for general studies of methods of discourse studies, see Titscher, Meyer, Wodak, & Vetter, 2000;Wodak & Meyer, 2015). This chapter will be limited to only some methods of qualitative discourse analysis, also since other methods are discussed in other chapters of this Handbook.
The general theoretical framework of this chapter, as well as of my other work, also on racist discourse, is a multidisciplinary approach called sociocognitive discourse studies, which links a systematic analysis of discourse structures, with an analysis of societal structures via a cognitive interface (for recent summaries, see e.g., Van Dijk, 2015a, 2015b. My studies on discourse and racism are part of a broad academic perspective of Critical Discourse Studies, focusing on the study of discursive domination and discursive resistance against domination (Fairclough, 1989(Fairclough, , 1995Van Dijk, 2008b;Wodak & Meyer, 2015).

What is Discourse?
Outside of the field of Discourse Studies, it is often asked to define discourse. Such a definition, however, is implicitly provided by all theories of the field, studying the many properties of discourse. In the same way, sociology does not provide a definition of society, but only describes and analyzes a myriad of properties of societies. A summary of these different properties of discourse may however be taken as a complex 'definition': • Discourse is a form and unit of language use.
o Discourse is an ordered sequence of words, sentences or turnseach with its own structures ▪ These sequences express coherent sequences of local and global meanings • The performance of these meaningful sequences in a communicative social context accomplishes speech acts and other forms of social action. • Discourse is form of communication • Discourse is a multimodal message (spoken, written, images, sounds, music, gestures) • Discourse is a form of social interaction • Discourse may instantiate a social relation, such as power, domination or resistance.
• Discourse may be a political action, such as a speech in parliament or party propaganda • Discourse is a cultural phenomenon, such as a conference paper or editorial • Discourse is an economic commodity or resource (it may be bought or sold, e.g. as newspaper or book, or represent a form of power).
Each of these very general definitions (and there are more) may be further specified, as we did for the definition of discourse and a form of language use. Thus, discourse as social interaction may be further defined in terms of speech acts such as promises and threats, or in terms of the many structures of conversation, such as Openings, Closings or Interruptions.
There are many genres of discourse, such as (many sorts of) informal conversation, news reports in the press or on television, parliamentary debates, party propaganda, many types of legal (laws, interrogations), political (speeches of politicians) or educational (textbooks, classroom interaction) discourse, advertisements, Twitter or Facebook messages, and so on. As we shall see in more detail below, these genres are defined in terms of the properties of the communicative situation or context, such as Who, When, Where for Whom and How the discourse is used, as well as by their style or meanings.
These genres may also be described in terms of the type of discourse structure, such as although both are forms of social interaction. We shall come back to these different properties of discourse and discourse genres below.
Since discourse can be defined in as many ways as shown above, Discourse Studies as a field is typically multidisciplinary, and not limited to more 'linguistic' approaches.
Conversation Analysis, for instance, is based on a more general, sociological study of interaction. And Critical Discourse Analysis focuses on the study of the discursive abuse of power (as in racist or sexist discourse) or the resistance against such domination (e.g., by antiracist or feminist discourse) also studied in political science. Ethnographic approaches may combine a study of discourse genres and their style, with complex analyses of the sociocultural situation, who (may, must) speak, when, to whom, on what occasion, in what circumstances and with what social or political functions and effects.
The notion of discourse is sometimes also used in a more abstract, and hence vaguer sense, for instance as a general philosophical or political system, for instance as the discourse of modernitysometimes written with a capital D: Discourse (Gee, 1999).
Philosophical approaches to discourse usually deal more with ideas (as expressed in discourse) than with the detailed structures of text and talk (for instance in the work of Foucault, 1980) and do not offer systematic methods for the analysis of such structures.
Finally, the term 'discourse' is not only used as a specific instance of text or talk, but also in a more generic sense, for instance as a class of discourses associated with a social context. For instance, 'political discourse' refers to a whole class of different genres of political discourse, such as parliamentary debates or party propaganda. Thus, below we speak of Migration Discourse, and thereby refer to a large class of all discourse genres of or about migrants or migration. If not in this generic sense of a class of discourse genres, in this chapter discourse is only defined and analyzed as a specific instance of language use, communication or interaction, for instance a specific news report, a specific debate in parliament, or a specific everyday conversation.

Migration Discourse
The complexity of discourse as a linguistic, social, political and cultural object or phenomenon also characterizes migration discourse, which represents a vast class of different discourse genres. The class of these genres is primarily defined in terms of their reference, that is, what they are about: the many aspects of migration as a social and political phenomenon. Other general classes of genres of discourse are for instance political discourse, media discourse or educational discourse.
News reports, editorials, parliamentary debates, laws, or everyday conversations are among the many discourse genres that may be about migration in general, and related phenomena, such as migrants (Them), autochthonous peoples (Us), causes of migration, integration, xenophobia, discrimination, racism, immigration policies, and so on, in particular.
Migration discourse not only may be about migration or its many aspects, but also be constituent part of migration as a phenomenon, as would be the stories of migrants, as well as parliamentary discourse preparing immigration policies. Contemporary discourse studies emphasize this fact, namely that discourse is not just a form of language use, but also a form of social and political (inter) action. Thus, migration as a social phenomenon not only consists of (groups of) participants, institutions, many types of social and political (inter)action, but also, quite prominently, of many genres of migration discourse as social and political acts and interaction.
After the more general introduction about Discourse Studies as a field, as well as its objects and methods, and a brief general characterization of Migration Discourse as a class of discourse genres, let us now proceed with a more systematic method for the study of migration discourse.

Genres and Contexts of Migration Discourse
There are many ways to engage in qualitative discourse analysis. Specialized studies may directly focus on just one aspect of discourse, for instance the structures and uses of pronouns, metaphors or argumentation. A more general methodology for the study of migration discourse, as presented here, proposes an overall, systematic introduction to various levels and dimensions of discourse structure and its uses and functions in the social and political context. This means that only a few of these structures, among hundreds of others, can be dealt with here.
One of the first theoretical and methodological tasks when studying (migration and other) discourse is to establish what genre of discourse one is analyzing. Above, we already mentioned some of these genres, such as stories of migrants or parliamentary debates. But obviously there are many more genres of migration discourse, such as: • Media Discourse: News reports (press, TV, radio, internet); editorials; interviews; reportages; cartoons; letters to the Editor.
• Personal discourse: Everyday face-to-face conversations; letters, e-mail messages; internet participation (Facebook, Twitter, Chats) Since there are many dozens of discourse genres of and about migrants, in this chapter our examples will be limited to only a few genres, such as laws, parliamentary debates, news reports and textbooks. By way of example we have chosen some of the discourses related to the U.K. Immigration Act of 2016.

Contexts
To define the discourse genres mentioned above, one may distinguish between contextual and textual characterizations. The latter were more common in traditional discourse and literary analyses and shall be dealt with below. Today, it is recognized that important distinctive features of genres must be in terms of the communicative situation, because discourse genres are first of all a type of social activity. Such communicative situations may be characterized by a few main and secondary categories, as follows: A. Time/Period B. Place/Space/Environment/Institution C. Participants i. Social Identities, e.g., ethnic identities, origin, etc. ii.
Social Relations, e.g. of domination (power abuse), cooperation, resistance D. Ongoing (Inter) Action i. Ongoing discourse ii.
Other social acts (e.g., cooperation, protests, etc.) E. Personal and Social Cognition i. Goals of the current interaction ii.
Shared and Mutual Generic Knowledge (Common ground) of participants iii.
Shared social attitudes and ideologies about migration With this schema, a first definition of a discourse genre can be given. For instance, a parliamentary debate takes place in Parliament as an Institution, usually in a parliament building, at a specific date, day and time, during so many minutes, hours or days, with participants in their role as politicians, MPs and members of political parties, and with national, ethnic or gender identities, representing voters, participating in a complex interactional discourse (a debate), controlled by the current Speaker/Chair, in various speaking roles, performing various speech acts (assertions, questions, accusations) as well as several social and political acts, such as governing, legislating, representing the peopleand possibly defending or discriminating against minority groupswith several social and political goals, e.g. to limit immigration.
Such discourse and social acts are based on shared knowledge about migration, as well as attitudes and ideologies about migration or minorities. The content/meaning and structures of such parliamentary debates, as well as of other migration discourses, is to be specified separately, as proposed below, but it will generally be about some migration aspect or event. A similar contextual definition may be given for any of the other migration discourse genres.
For participants to write or speak appropriately in these variable communicative situations, they must adapt their text or talk to this situation, and they can only do so when they know, that is, mentally represent what is currently relevant in such a situation. Hence, contexts as they influence discourse are a special kind of personal, mental model, a context modelwhich may be slightly different for each speaker, so that also misunderstandings may arise (for details about contexts, context models and contextual analyses, see Van Dijk, 2008aDijk, , 2009). This context model exercises the overall control in the production of discourse, and makes sure it is appropriate in the ongoing communicative situation.
The relevance of such a contextual analysis of genres of migration discourse is shown by the fact that the discourses in many ways show or manifest aspects of the

Contextual Racism
As we shall see below, at all levels of discourse, but also at the level context, structures may be expressing or functioning as a form of racism. Thus, the same discourse theme, such as 'problems of immigration' may make a parliamentary speech (more or less) racist depending on the identities, relations and goals of the participants, for instance with the goal to limit immigration or to help migrants, or spoken by members of a progressive, proimmigration party or by an MP of an extreme rightwing party.

Themes and Topics
Once established the genre and the more detailed contextual parameters of a migration discourse, we may focus on the various structures of text or talk itself. We do so first by an analysis of the themes and topics of discourse, that is, their general, overall meaning.
Topics are theoretically described as semantic macrostructures, consisting of macropropositions that semantically subsume and control the lower level meanings (propositions) of the sentences of the discourse (Van Dijk, 1980). A practical way to define these macropropositions is to summarize a paragraph or whole discourse in the form of a few simple sentences. Such topics are typically expressed in the headline of a news report or the title and abstract of a scholarly paper, but we also may do so at the beginning of a story or conversation: "Yesterday I had an accident and…".
Topics not only are fundamental because they subsume the local meanings of a discourse, but also because they are the information that is generally best remembered by the recipients. We may forget the details of a news report, but have better memory for the overall topics. If a news report is about an immigration event, for instance, many refugees drowning when crossing the Mediterranean, we may later forget the details, or how many refugees drowned, but may remember the most important information, that is, the information at the top of the mental model we construe of the event by interpreting the news report (Van Dijk & Kintsch, 1983). In other words, macrostructures of discourse may correspond with macrostructures of mental models, and these may be used as a condition for social action, including discriminatory practices.
Topics also allow a methodological short-cut when one wants to analyze a large number of discourses about migration, such as news articles or parliamentary speeches.
Instead of analyzing the complex local structures (see below) of such discourses, we may study only their topics by summarize these discourses in a few sentences. These may feature such categories as Time, Place, Participants in various identities and roles, as well as the global action or event (as we did for the structure of context above). Although such summaries do not provide all details, they at least provide a global idea of the meaning of a larger corpus of text or talk.
Although there is no explicit theory of themes, as there is for topics defined as semantic macrostructures, themes are concepts that define classes of discourses. Thus, all discourse genres we have mentioned above share the overall theme of migration. Or they may be about refugees, about the current elections or about the racism of the extreme right.
Whereas a topic, as defined, characterizes a specific discourse, as is the case for its expression in the headline of a news report, a theme may define a large number or sequences of news reports. Thus, each concept of a macroproposition (e.g., refugees, arriving by boat, drowning, etc.) may constitute a theme of many articles. In everyday practice, we use themes to refer to what a discourse is about, e.g., refugees, the elections, or racism. In research, theme-concepts may be used as a criterion for the selection of discourses for a corpus, because they are about the same phenomenon.
Here are a few examples of topics as expressed in titles, headlines and summaries of migration discourses: (1) Immigration Act 2016 An Act to make provision about the law on immigration and asylum; to make provision about access to services, facilities, licences and work by reference to immigration status; to make provision about the enforcement of certain legislation relating to the labour market; to make provision about language requirements for public sector workers; to make provision about fees for passports and civil registration; and for connected purposes (Initial Summary of the U.K. Immigration Act of May 12, 2016). This summary of the U.K. 2016 Immigration Act condenses a bill of 236 pages (nearly 90,000 words) into a summary of 66 words, e.g., with such very general concepts as "access to services, facilities, licenses and work". Human Rights News summarizes some "key changes" of the same law in "plain English": (2) • Employers who hire illegal migrants and the workers themselves face criminal sanctions.
• Migrants who do not have permission to be in the UK can have certain privileges revoked. For example, their bank accounts can be frozen and their driver's license can be seized. • It will soon be a criminal offence for a landlord to knowingly rent premises to an illegal migrant. If found guilty, the landlord can face up to five years in prison. This law will take effect when the Secretary of State creates regulations that state the law's 'start date' (http://rightsinfo.org/immigration-act-2016-plain-english)(May 31,2016).
We see that a summary (of a law, news, etc.) may very well be subjective, for instance as a critical commentwhich focuses on the social consequences of the new law (criminal sanctions, etc.).
Crucial is that such summaries, rather than detailed text, generally are best remembered by the readers. See also the following headline and summary of a news report in the Guardian: (3)

Immigration to UK hit record levels prior to Brexit vote, data shows
Record level of immigration, at 650,000 people, driven by a historically high inflow of 84,000 EU citizens before referendum. Immigration to Britain reached 650,000its highest ever annual levelin the runup to the EU referendum, fuelled by record numbers of European migrants coming to work in "the jobs factory of Europe", official figures reveal (Guardian, 1-12-16).
We see that whereas the headline uses the more general notion of "record levels", the first (topical) sentence (and the rest) of the news reports specify the more precise numbers.
Incidentally, news about immigration is replete with numbers (e.g. of arrivals), which may also have a rhetorical function, when emphasizing the (vast) numbers of immigrants, a rhetorical ploy called the "Numbers Game" (Van Dijk, 1991; and see below).
The next day, on December 2, 2016, the Sun published an editorial with the following headline about the same theme (a Government report on immigration statistics): (4) Shocking immigration stats reveal we have no control of our bordersand the government must do more to cut down numbers (Sun, This is no longer a summary of the facts, as in the news, but a summary of the Sun's opinion about immigration. Apart from the strong appraisal word "shocking", notice the presupposition (see below) that we have no control of our borders, and the deontic modality (see below) in the recommendation of what the government "must" do.

Schematic Superstructures
Whereas topics or semantic macrostructures define the overall meaning of a discourse, discourses also have an overall superstructure, a schematic form or format. Some of the categories that constitute such superstructures have already been mentioned above such as the Title, Abstract or Conclusions of a scholarly article, or the Headline and Lead of a news report, as we have seen for the Guardian and Sun. These schematic categories define the overall organization of text or talk, for instance in what order its components occur, and which components are part of higher level components. For instance, the Headline and Lead of a news report are both part of the initial Summary of the news report.
Superstructures are often conventional for a discourse genre, as is the case for most news reports or scholarly articles. Even informal conversations may have some more or less conventional schematic categories, such as Openings (e.g. by Greetings:. Hi) or Closings (by leave-taking: Bye!).
Very complex texts, such as the U.K. Immigration Act is structured by nine Parts, each consisting of several Chapters (such as Chapter 2, "Illegal Working") which again have sub-categories (such as "Offences")all with summarizing titles.
There are superstructures that play a fundamental role in the organization of everyday discourse genres, as is the case for narrative structures of storytelling, or the argumentative structure of debates, editorials or other opinion discourse.
For instance, the classical narrative structure of an everyday story may be summarized with the following categories: Orientation, Complication, Resolution, Coda/Conclusion roughly in that order, with an Evaluation (such as 'I was so afraid') which may occur anywhere in the story (Labov, 2013). Similarly, an argumentation typically consists of different types of general or specific Premises, followed (and sometimes preceded) by a Conclusion (for detail about the structures of argumentation, see, e.g., Van Eemeren, 2014).
Schematic superstructures characterize types of discourse structure or whole genres more or less independently of their meaning. This means that methodologically they may appear to be less relevant for research focusing on content. Thus, all MPs in a parliamentary debate make use of argumentation structuresincluding fallacieswhether they are in favor or against immigration or any other aspect of migration. Being in favor of or against immigration, thus, is an element of meaning, not of form. In other words, superstructures of discourse, as such, may not be racist or antiracist -only their 'contents'.
However, because superstructures are about overall order and organization, they may make their global meaning (topics) more or less salient: As we have seen above, the meaning of a headline may represent the top of the macrostructure of a text and of a mental model of the text, and hence tends to be better recalled. The same is true for Abstracts and Conclusions of discourse genres. More generally, superstructures organize discourse segments, and hence organize the macropropositions (topics) that subsume the meaning of such segments.
In empirical research of migration discourse it is thus not only important to establish overall meanings but also their order to position in discourse. Indeed, many readers may only read (and hence remember the information expressed by) the Headline or Leads of news reports, or the Title, Summary or Conclusions of scholarly articles.

Local Meanings
Besides the study of the overall meanings (topics, themes) of migration, most relevant for qualitative discourse analysis is the study of its 'local' meanings, that is the meanings of words, sentences or sequences of sentences.
Traditionally, both global and local meanings are represented as propositions, consisting of a number of arguments referring to things or people, and a predicate, referring to a property, relation, action or event, as in the proposition 'The voters elected a right-wing president', where 'voters' and 'president' are arguments, and 'elected' is the main predicate, and 'rightwing' a lower level predicate modifying 'president.
Today, such propositional meanings are also represented as more complex conceptual schemas. For instance, in the proposition just mentioned, the argument 'voters' should be inserted in a schematic category Agent, whereas 'president' would be part of the schematic category Patient, and 'elected' part of a category Action. Probably, the underlying mental models that represent the meaning of a sentence or whole discourse, or the situation it is about, maybe organized in terms of such a schema. Similarly, at the level of words, one may also use schemas to represent individual word meanings or concepts. For instance, the concept of 'president' is a complex schema consisting of various categories defining 'politician', 'head of state', etc. (for detail, see studies of cognitive linguistics, e.g., Croft & Cruse, 2004;Hart & Lukes, 2007).
Observe that this propositional meaning is independent of the syntactic form of the sentence, which might also be expressed by a passive sentence A rightwing president was elected by the voters. As is the case for superstructure categories, also the syntactic structure of a sentence may make specific meaning elements more salient, indicate whether some aspect of the meaning is already known or expected (the 'focus' or the sentence), or whether something new or unexpected is being communicated (the 'comment' of the sentence). In this way one might hide or mitigate the negative actions of dominant groups, for instance in headlines such as "Black student killed by police," or even simply "Black student killed," as has been found in many studies of racist discourse (Van Dijk, 1991, 1993; see also the first study of critical linguistics : Fowler, et al., 1979).
As a concrete example of local meaning, see for instance the following paragraph consisting of two sentences in the Guardian article mentioned above: (4) The immigration minister, Robert Goodwill, responded to the figures by saying the British people had sent a very clear message that they wanted more control of immigration. He renewed the government's commitment to getting net migration down to sustainable levels in the tens of thousands (Guardian, 1-12-16).
The first sentence consists of various clauses expressing a hierarchical structure of underlying propositions about the response of the minister, talking about the message of the people and the content of that message, with the minister and the (cited) people as Agents of various actions. Obviously, such a first semantic analysis in terms of propositions and their structures (which itself needs to be refined) requires further critical analysis, e.g., by describing the minister as pretending to speak for "the people," and interpreting the numbers in terms of the Conservative Governments policy to limit immigration. Other such semantic aspects of this article and other texts will be mentioned below.
With these basic notions of semantics, we may now proceed to study a potentially large number of meaning aspects of discourse, only some of which will be briefly summarized here because they are offer methods for relevant qualitative discourse analysis.

Modalities
Propositions, representing meaning, can be modified in many ways, typically so by different kinds of modality, such as those of necessity and possibility (alethic modalities), or what is obligatory or permitted (deontic modalities). For instance, alethic modalities may be expressed by adverbs such as probably, or auxiliary verbs such as may, might or must. Thus, the proposition about the election of a rightwing president may be expressed in modal sentences such as Probably the voters will vote for a rightwing president, or The voters may elect a rightwing president. Similarly, deontic modalities may be expressed in such sentences as The refugees must leave the country or in The refugees may stay in the country. In the Guardian article, we find for instance the following example: (5) Some early data from after the Brexit vote suggest that some migration to the UK may be decreasing (Guardian, 1-12-16).
Both the verb "suggest" as well as the modal auxiliary "may" express that the conclusion drawn from the numbers is not certain (according to the Guardian). In a more critical perspective, one may find that other newspapers, typically the rightwing tabloids, tend to

Implications
Propositions may be explicitly expressed by sentences and discourse, but also remain implicit. Such implications can nevertheless be construed by the recipients by applying their shared knowledge of the world. A political discourse may explicitly say that refugees may not stay in the country, and recipients may then infer that refugees must leave the country, even when that is not explicitly said.
More generally, because of such plausible inferences based on world knowledge, the interpretation of a discourse is not limited to the propositions that are explicitly expressed but also includes the most plausible implications that can be derived from the explicitly expressed propositions. In more cognitive terms, we may say that the complete, subjective meanings of a discourse as assigned by speakers or recipients, is in their mental model of an event or situation. By general pragmatic rule, only some of the propositions (schemas, concepts) of such a mental model need to be expressed in discourse, namely the information that is as yet unknown and cannot be inferred by the recipients.
Methodologically, implications are important in semantic discourse analysis because they represent meanings that are construed in their mental models by all language users with the same world knowledge, that is, the members of the same epistemic community, but speakers may claim they never actually said what was implied. That is, implicit meanings can be denieda deniability that may be crucial in many (6) If we are to continue building an immigration system that is fair to British citizens and people who come here legitimately to play by the rules and contribute to our society, we must ensure that it is balanced and sustainable, and that net migration can be managed. When properly managed, immigration enriches this country, as we benefit from the skills, talent and entrepreneurial flair that people bring to our society. But, as I said in my recent speech, when net migration is too high, and the pace of change is too fast, it puts pressure on schools, hospitals, accommodation, transport and social services, and it can drive down wages for people on low incomes. So we must achieve the right balance, rejecting both extremes of the debate, from those who oppose immigration altogether to those who want entirely open borders. That is why, since 2010, we have worked to build an immigration system that works in the national interest, one that is fair to British taxpayers and legitimate migrants, and tough on those who flout the rules or abuse our hospitality as a nation (Teresa May, House of Commons, 13-10-16) On superficial reading, such a discourse expressing positive opinions and attitudes about immigration ("immigration enriches the country," etc.) and about the governments immigration policies and laws ("fair", "balanced"), are typical forms of positive selfpresentation of much ideological discourse. But an expression such as "people who come here legitimately" politically implies that illegal immigrants do not contribute to our societyand that the law will have provisions to punish them. The metaphor "who play by the rules" has the same implicationsrepresenting immigration as a game between government, citizens and immigrants. Similarly, "when properly managed, immigration enriches this country", implies that if not properly managed (implying as provided by the law) immigration does not enrich the country.
The first sentences, thus, function as the first part of a disclaimer, beginning with positive self-presentation, followed by but in line (5) introducing all the negative aspects of badly managed (implying: controlled) immigration. These negative consequences themselves are however formulated in terms of euphemisms "pressure on schools, hospitals, …" but imply what most MPs and other citizens will understand by its political implications: less place for Us in schools, hospitals, etc.information that is the typical second part of the disclaimer: a negative representation of Them.

Presuppositions
Another fundamental aspect of the meaning of discourse are its presuppositions, that is, propositions that are assumed to be true in order for another proposition to be meaningful.
More generally, presuppositions represent the knowledge shared by speaker and recipients (often called Common Ground) and which is relevant for the production and interpretation of a (fragment of) discourse. For instance, in the example of the election of a rightwing president, it is presupposed that there is a president and that there are votersand such presuppositions may be marked in the text by the definite article the. More generally, linguistic presuppositions are those propositions that are 'triggered' by specific discourse structures, such as factive verbs (to realize, to discover, to regret, to stop or to continue), some adverbs (such as even and also) or the position of clauses, e.g., initial that-clauses, as in the sentence That the voters elected a rightwing president, has serious political consequences, where the initial that-clause expresses (a proposition referring to) a known fact.
As is the case for implications, also presuppositions may be used to manipulate the interpretations (mental models) of the recipients, typically so in media and political discourse, which may refer to 'the criminality' of migrants, thus presupposing that migrants are criminal. More generally, thus, presuppositions in discourse may be used to manipulate the knowledge and other beliefs of recipientsthey signal that some fact is known to be the case, even when it is not. That is, they may function as oblique assertions, that is assertions that are not explicitly made, but indirectly.
In many discourses such presuppositions are innocent, e.g., when we mention "the waiter" in a story about a restaurant, we presuppose that there is a waiter on the basis of our knowledge about restaurants. Similarly, we may routinely talk about our father, apartment or car, without first asserting that we have a father, apartment or car.
In the example of Home Secretary May in the debate of the UK Immigration Bill, we find such a presupposition (a'), triggered by the verb "to continue" in sentence (a): (a) If we are to continue building an immigration system that is fair to British citizens (a') We are building an immigration system that is fair to British citizens The paragraph following the paragraph of Teresa May we have cited, begins as follows: (8) Over the past five years we have taken firm action to reform the chaotic and uncontrolled immigration system we inherited, and to ensure that people are coming here for the right reasons.
This sentence presupposes that we (the Conservatives) inherited a chaotic and uncontrolled immigration (from Labour). Notice that she does not explicitly assert this about Labour's system, but presupposes it, as a generally known fact or shared opiniona typical example of oblique assertion by presupposition. This is typical for assertions that are controversial and can thus be hidden or made less prominent when expressed as a presupposition.

Actor and Action Descriptions
Migration discourse typically mentions "migrants," as well as "ourselves," "our country," "the nation", "taxpayers" or "British citizens", as we have seen in the debate fragment of Teresa May above. One of the prominent properties described in the semantics of discourse, is the way the actors or participants are referred to and described. Indeed, such discourse may mention "migrants," "immigrants," "refugees" or "foreigners", in general, or "Syrians" or people of other nationalities, in particular. The references may be generic (all migrants) or specific (a particular migrant), whether or not with proper names. Further analysis of the way actors of events are described may distinguish between actors by their functions (MPs), their membership of groups, their nationality, their gender, age, profession, and so on (for a detailed system of actor description, see *Van Leeuwen, 1996).
A systematic analysis of the way actors are described in immigration discourse, provides insight in underlying attitudes about immigrants, e.g., when U.K. tabloids used to describe refugees as "scroungers" who abuse of welfare provisions (Van Dijk, 1991). The same is true for the description of their actions, for instance when Teresa May metaphorically describes them as "flouting the rules" or as "abusing our hospitality" (a description that itself presupposes that "we" are hospitable).
These qualitative and critical analyses of actors (Us vs. Them) in migration discourse may be further elaborated by a more quantitative approach using frequencies of occurrence, as was also typical in traditional Content Analysis. Today, such a more quantitative analysis is provided by various methods of Corpus Linguistics (see, e.g. Baker 2012), in which we may also show how and how often a word co-occurs with other words in the same data lines of a corpus, for instance if the word "immigrant" typically co-occurs with "illegal".

Further semantic analysis
We have mentioned only a few properties of discourse typically studied by a semantic analysis of local and global meanings. Depending on the aims of a research project, the genre of discourse, the size of a corpus or the time or expertise of the researcher, there are many other aspects of meaning that may be studied. For instance, a study of metaphors (such as "waves" of refugees) may tell us something about the threatening mental models of migrant discourse. We may describe migration events at various levels of generality or specificity, higher or lower granularity (more or less vague or precise, with more or less details at each level), and so on. An epistemic analysis may focus on the way knowledge is expressed, implied or presupposed, as we have seen above (Van Dijk, 2014).

Ideological analysis
Especially a more critical study of migration discourse typically also engages in a study of the way ideologies are expressed. Ideologies are fundamental, socially shared mental representations of social groups: racists, anti-racists, pacifists, militarists, feminists, sexists, neoliberals, socialists, and so on. They are typically organized by polarization, (good) ingroups vs. (bad) outgroups, a polarization that also be expressed in discourse, e.g., between (good) Us vs. (bad) (Them), as we have seen in the examples above. Ideologies represent the characteristic identity, actions, aims, norms and values of a group, and may control more specific attitudes, e.g. about immigration, integration or adaptation of migrants (or other attitudes, such as abortion or the death penalty). These more specific attitudes in turn may influence the personal mental models of (e.g. migration) events of the members of an ideological group. And depending on the communicative context of a discourse, these ideologically based (and biased) mental models may finally influence the way discourse about such events is expressed. We see that between fundamental ideologies (e.g. of racism) and actual racist text or talk, there are various levels of sociocognitive analysis.

Conclusion
Migration is a complex sociopolitical phenomenon that has been studied in most of the humanities and social sciences. One fundamental way to study migration is to analyze the properties of the many forms of text or talk of or about migrants. The multidisciplinary field of Discourse Studies, prevalent in all the humanities and social sciences since the 1960-1970s, offers sophisticated theoretical and methodological frameworks for a systematic and explicit study of migration discourse. Beyond traditional Content Analysis, and more explicit than popular but methodologically vague Frame Analysis today, the quantitative and especially the qualitative methods of contemporary Discourse Studies offer insight in the many ways migration discourse is structured, how it expresses