Manufacturing dissent: The discursive formation of nuclear proliferation (2006–2012)

Introduction

After a glimmer of optimism at the turn of the century, the 21st century is once again emerging as one of continuing crisis and conflict (Houtart, 2010). One way in which the sense of global crisis has intensified over the past decade is through the proliferation of nuclear weapons which the United Nations (UN), as well as the US and the UK media, maintain took place in Iran and North Korea over this period. In this article, we examine the ways in which nuclear proliferation was discursively constituted between 2006 and 2012. Techniques of discourse analysis and corpus analysis are combined to compare resolutions passed by the UN Security Council (UNSC), and articles in prominent broadsheets from the United States and the United Kingdom, two English-speaking permanent members of the UNSC. Specifically, we will undertake three lines of enquiry: (a) the ways in which nuclear proliferation was discursively constituted in the international political sphere (2006–2012), (b) the ways in which nuclear proliferation was discursively constituted in the public sphere in the United Kingdom and United States (2006–2012), and (c) the ‘contradictions’, ‘changes’ and ‘transformations’ (after Foucault, 1972) that took place as the discourse of nuclear proliferation was delocated from the political sphere and relocated in the public sphere.

The discursive formation of nuclear proliferation

For some time now, it has been proposed that any discrete body of knowledge is constructed not as a set of self-evident truths made manifest through the transparency of observable data, but rather is characterised by the particular set of objects, subjects, concepts and strategies which are constituted within a distinctive ‘formation of discourse’ (Foucault, 1972). One way of understanding the distinctiveness of a discursive formation is to analyse the ways in which words, statements and texts combine systematically to bring a particular view of the world into being: ‘to define a system of formation in its specific individuality is … to characterize a discourse or a group of statements by the regularity of its practice’ (Foucault, 1972: 74). Arguably, no area of knowledge or interhuman engagement is more contingent upon discursive practice than international relations. Here, sets of intergovernmental legal, political and economic associations are created, maintained and transmitted across two realms: a political sphere located in national executives and intergovernmental organisations such as the UN, and a public sphere informed by a national press alongside electronic media such as television, the Internet and social networking. It is a key principle of this article that the distinctive features of a particular nexus of international events such as ‘nuclear proliferation’ are constituted not within any one particular set of texts, but in the dynamic relations that operate between different text types which circulate within and between different institutional sites, both national and international.

The analysis of the discourse of nuclear proliferation goes back to speeches and documents from the ‘Cold War’, the long running nuclear stand-off which took place between the United States and the Union of Soviet Socialist Republics (USSR) between 1947 and 1991. Using a heterogeneous approach, Mehan, Nathanson and Skelly (1990) concluded that the Cold War was constituted in the interstices between documents generated by the US administration, the National Council of Catholic Bishops, the Roman Catholic Church and the speeches of Ronald Reagan. This ‘intertwined system of discourse’ (p. 158) had material consequences. For the United States and its allies, the ‘dialogic’ process which took place between ‘strategic analysts’, ‘the Catholic Church’ and ‘peace groups’ (p. 135) gave rise to three discursive ‘strands’ – the threat of Soviet expansion, reliance on nuclear weapons to curb the threat from the USSR and the role of nuclear weapons – which ultimately served to deter global warfare over this period.

In studies addressing media texts more similar to our own, critical discourse analysis (CDA) has been used to consider the cross-cultural positioning of Iran within the discourse of nuclear crisis. Izadi and Saghaye-Biria’s (2007) analysis of editorials relating to the Iran nuclear crisis in The New York Times, The Washington Post and The Wall Street Journal from 1984 to 2004 revealed a range of lexis which presented Islam as a threat and Iran’s government as untrustworthy, while failing to acknowledge the United States’ refusal, along with other signatories, to realise their own commitment to the Nuclear Proliferation Treaty (NPT, 160–161). Using a similar analytical framework, Behnam and Zenouz (2008) undertook a comparative analysis of articles relating to the Iran nuclear crisis from a corpus of Iranian and British broadsheets published in 2004: the conservative Kayhan and Telegraph, and the more progressive Iran Daily and Guardian. All four newspapers portrayed Iran as salient within the systemic-functional transitivity system, being realised either as an Actor, Senser or Sayer. However, papers on the left of the political spectrum in both countries gave particular prominence to verbal process types, acknowledging that the positions described are indeed ideological. One implication of this predominance of verbal processes in the Iranian papers is that Iran’s nuclear policy appeared to be supported by a political consensus within the country. By contrast, right wing newspapers in both countries tended to describe events using material processes, which bestowed them with greater facticity. British newspapers presented a dichotomised worldview, where the European Union (EU) and Iran occupied polarised positions, with Iran’s uranium enrichment programme coming in for particular vilification (pp. 213–216).

Most recently, Rasti and Sahragard (2012) analysed the patterns of discourse in 23 articles on Iran’s nuclear programme published in the Economist between 2007 and 2010. Once again they described a polarisation at the lexical level between Iran and the West, where the western ‘we’ was portrayed positively as supporting sanctions, while an Iranian ‘they’ was represented in a negative light, particularly with regard to ‘confidence-building activities’ in relation to arguments around the lifting of sanctions (pp. 735–736). This dichotomisation was further heightened through the use of a range of conceptual metaphors within the periodical (pp. 737–739). For the most part, the reactions of Iranian citizens were elided from the presentation of the arguments around sanctions and, where they did occur, were represented as edgy, unpredictable and self-interested (p. 740). At the level of discourse, temporality was frequently invoked to legitimise a negative orientation towards Iran and its nuclear programme along with invocations to an ‘indeterminate’ group of international actors to position themselves. Additionally, a range of argumentation strategies were used to legitimate international action against Iran, including the topoi of ‘usefulness/advantage’, ‘danger’, ‘justice’ and ‘responsibility’ (pp. 743–744). Finally, a negative perspective on the Iranian nuclear issue was generated by using nominalisation and passivisation to elide attribution of agency, as well as referring to many more Western than Iranian sources (pp. 745–746).

Discourse studies relating to nuclear proliferation in North Korea are less plentiful than those in relation to Iran. An early paper (Min, 1999) undertook a critical-linguistic, cross-cultural discourse analysis of 92 articles harvested from The New York Times and the Korea Herald relating to the first four months’ coverage of the nuclear stand-off between North Korea and South Korea in 1994. While headlines (or ‘macro propositions’) in The New York Times accorded the US prominence in mediating the North Korean nuclear talks (p. 8), headlines in the Korea Herald focused more locally on North Korean transparency for the security of the Korean peninsula and the freezing of North Korean nuclear activities. The New York Times positioned US participants as the agents in the negotiating processes, while the agency of the Korean participants was downplayed through a variety of rhetorical devices (pp. 11–12). Unsurprisingly, the Korea Herald increasingly constructed the South Korean participants as principal agents in the US–North Korean nuclear talks, with the United States being accorded agency in a less prominent fashion (p. 14). This was realised in part through the respective thematisation of the United States in the New York Times headlines (pp. 13–14) and South Korea in the Korea Herald headlines (pp. 23–24). With regard to lexicalisation, in The New York Times there was a considerable degree of over-wording around the vocabulary related to the notion of crisis, with the use of associated words such as ‘crackdown’, ‘breakthrough’, ‘curbs’ and the word ‘crisis’ itself (pp. 10–11). By contrast, the Korea Herald did not exhibit such plentiful lexicalisation around the theme of crisis; in fact, it attested to portraying events on the Korean peninsula as generally less unstable (pp. 19–21).

While the studies of the discourse of nuclear proliferation crisis in Iran and Korea reviewed above have yielded some interesting insights into the ways in which lexis is used for ideological effect, they have all restricted themselves to analysing just one type of media text. Furthermore, the corpus size has necessarily been small due to their qualitative approach. While Mehan et al.’s approach is perhaps closest to our own study inasmuch as they do consider the relations between different text types, their corpora remain relatively small and subject to manual analysis. By contrast, our study uses corpus analysis techniques to engage critically with the constitution of the nuclear proliferation across different text types, as a discursive formation (after Foucault, 1972). In particular, we aim to problematise the ways in which the notion of ‘crisis’ is constituted as texts are delocated from the political sphere and relocated in the public sphere.

Methodology

The methodological approach we take in this article combines elements of poststructuralist discourse theory (after Foucault, 1972), critical linguistics (after Kress and Hodge, 1979) and corpus linguistics (after Stubbs, 1996). While the individual methodological strands we draw on are well established in communication studies and applied linguistics, it is rare for the conceptualisation of discursive formation to be analysed using corpus techniques, and to the best of our knowledge only Baker and McEnery’s (2005) analysis of articles about refugees and asylum seekers published in UK newspapers and by the Office of the United Nations High Commissioner for Refugees (UNHCR) has attempted to look at the linguistic and discursive relations between texts across the political and the public spheres. Our study arises from a multidisciplinary research project carried out at the University of Warwick, United Kingdom, that investigates the dynamics that shape the emergence, success and failure of crisis leadership in global economic and security governance (CLiGG). ¹

Our corpora were compiled from two different sites: the UNSC and prominent national press outlets in the United Kingdom and the United States, commonly regarded as two of the most powerful permanent members of the UNSC. The UNSC resolutions corpus (Table 1) comprised all UNSC resolutions between 2006 and 2012 that were available online (http://www.un.org/en/sc/documents/resolutions/). Between 2006 and 2012, there was a range of 48–88 resolutions per year, with an annual average of 63 resolutions and an average length of 1558 words per resolution.

OPEN IN VIEWER

Table 1.

UNSC resolutions corpus.

	Words	Texts
2006	78,981	88
2007	75,074	57
2008	88,817	65
2009	87,650	48
2010	95,478	59
2011	114,085	66
2012	103,566	53
2013	7041	4
TOTAL	650,692	440

Four newspapers were selected for analysis: The Times and The Guardian from the United Kingdom, and The New York Times and The Washington Post from the United States. Not only do these papers from each country represent a spread of political viewpoints, from the ‘left’ to the right’ of the political spectrum, but they also are ‘broadsheets’ which are conventionally regarded as publishing more verifiable, ‘serious-minded’ content than – for example – UK tabloids. The US/UK broadsheets corpus (Table 2) comprised all articles containing the phrase ‘nuclear crisis’ or ‘nuclear proliferation’ between the years 2006 and 2012 from these four papers. The data were collected using the news database site Nexis UK. Between 2006 and 2012, the broadsheets produced 162–361 nuclear-related articles per year between them, with an average of 227 articles per year. The documents in the corpus were converted from their original, varied formats to a uniform text format amenable to machine analysis and the corpus was cleaned by inserting header, footer and body tags so that the considerable metadata could be filtered from any corpus analysis.

OPEN IN VIEWER

Table 2.

US/UK broadsheets corpus.

	Newspaper	Words	Texts
Total 2006	Guardian	37,798	51
	New York Times	118,822	105
	Times	33,059	46
	Washington Post	74,008	72
Total		263,687	274
2007	Guardian	31,429	46
	New York Times	76,966	63
	Times	18,772	25
	Washington Post	59,567	49
Total		186,734	183
2008	Guardian	20,158	30
	New York Times	79,321	61
	Times	16,830	23
	Washington Post	63,750	63
Total		180,059	177
2009	Guardian	52,700	61
	New York Times	62,949	59
	Times	53,314	69
	Washington Post	54,249	52
Total		223,212	241
2010	Guardian	24,589	32
	New York Times	54,984	55
	Times	42,129	56
	Washington Post	43,308	49
Total		165,010	192
2011	Guardian	52,616	79
	New York Times	64,588	50
	Times	65,115	115
	Washington Post	104,438	117
Total		289,757	361
2012	Guardian	25,577	29
	New York Times	63,608	66
	Times	10,161	19
	Washington Post	53,505	48
Total		152,851	162
TOTAL		1,461,310	1590

Corpus analysis techniques were used to ascertain the content of the corpora and the comparability of one corpus with another. First, the word list program in Wordsmith Tools (Version 5, Scott, 2008) was used to establish basic statistical information about each corpus and the overall frequencies of the lexis in each corpus. An analysis of lexical trends was then carried out using the keywords program to establish which words were salient in each corpus, using the log-likelihood algorithm (hereafter LL) in order to determine whether words appeared more or less often than might be expected by their observed frequency in either of the two corpora rather than the reference corpus (after Scott, 2008). Words were identified as ‘key’, where their difference in relative frequency was statistically significant when compared with the respective reference corpus (p < 0.000001, after Baker, 2006). Since each corpus comprised a different genre and was also of considerable difference in size, it was not appropriate to compare them directly. Not least, this would have resulted in much of the statistically significant lexis revealing differences between the resolution and the newspaper genres, but not necessarily the ways in which the ‘objects, subjects, concepts and strategies’ relating to nuclear proliferation were themselves constituted. Since it was also not possible within the scope of the project to create a massive comparator corpus of all the articles from the four broadsheets published between 2006 and 2012, the broadsheets corpus was compared separately with the British National Corpus (BNC, 2007) as a common, baseline ‘reference’ corpus (RC). ² However, it was possible to carry out the keyword analysis of the UNSC resolutions corpus with a greater level of precision by creating a sub-corpus of resolutions which specifically related to nuclear proliferation, as indicated by the document titles on the UNSC webpages. Between 2006 and 2012, there were 18 resolutions focusing on nuclear proliferation, ranging between one and four per year. This sub-corpus was then compared with the larger corpus which comprised all UNSC resolutions produced between 2006 and 2012.

Following Baker (2010), the top 100 keywords in each corpus (Appendix 1) were subjected to further qualitative checks and manual, context-sensitive interpretation (see also Baker, 2006: 125; Baker and McEnery, 2005). To do this, we used a combination of concordance and collocation data to reveal cross-corpus variations in linguistic phenomena identified previously in the sample. First, senses and roles realised by the keywords were checked via concordance, whereby all the occurrences of a particular search term in a corpus are presented within its linguistic context (Baker, 2006: 71). Concordance data were sorted for qualitative analysis; in particular, it was ordered alphabetically according to other salient words which occurred to the left and to the right of the keyword in question, in order to reveal regularities of co-occurrence within each corpus. Where it was necessary to understand the broader context of a term, concordance lines were expanded and longer sections of text were subjected to interpretative analysis. Second, data were examined relating to the collocation of keywords, that is to say their tendency to appear in combination with or in the company of other words. Patterns of collocation occurring five words to the left and right of each keyword were considered, and the most frequently occurring collocates were then selected in order to consider the semantic implications of their co-occurrence with the keyword in question. Finally, the linguistic data were grouped together under emergent themes relating to the three lines of enquiry set out at the beginning of the article. In the analysis that follows, keywords are set out in capital letters (e.g. ‘IRAN’) and collocates are set in italics (e.g. said).

A comparison of UNSC resolutions and US/UK broadsheets

A comparative analysis of the hundred strongest keywords in each corpus revealed that most salient lexical items fell into three categories signifying actors, strategic actions and technologies, which therefore emerged as the principal constituents of the discourse of ‘nuclear proliferation’. The majority of these lexical items – around half the strongest keywords in the newspaper corpus and around a third of those in the resolutions corpus – were constitutive of different actors, especially nation states and human agents. The following section reports on how these two prominent types of actor were constituted in the political and public spheres.

Actors

Twelve discrete nation states featured within the 100 strongest keywords within the newspaper corpus. Of these, the most prominent were IRAN, KOREA, UNITED STATES and CHINA. By contrast, only seven discrete nations occurred within the strongest 100 keywords in the UNSC resolutions, with IRAN and DPRK – the Democratic People’s Republic of Korea – appearing as the strongest. In the resolutions corpus, IRAN (n = 340) collocates with nuclear (n = 79), programme (n = 41), activities (n = 26) – often co-occurring in regularly recurring phrases describing ‘nuclear activities’. In particular, in the discourse of the UNSC resolutions, the noun ‘Iran’ is regularly embedded within lengthy noun phrases which are the butt of some measure of a censure against the country. For example, the idiosyncratic, technological phrase ‘Iran’s proliferation sensitive nuclear activities’ recurs eight times across five different resolutions, for example,

Calls upon all States to exercise vigilance regarding … individuals who are engaged in, directly associated with or providing support for Iran’s proliferation sensitive nuclear activities or for the development of nuclear weapon delivery systems.

Furthermore, compared with other countries Iran is positioned within the resolutions as the butt of UNSC actions, especially as ‘receiver’ of verbal processes (after Halliday and Matthiessen, 2006) as evidenced by the recurring phrases such as ‘persuade Iran …’ (n = 9), ‘calls upon Iran …’(n = 7), ‘encourages Iran …’(n = 6) and ‘urges Iran …’(n = 1), as well as the mental process ‘review Iran’s actions …’(n = 4). Of the instances where Iran is actually accorded agency in the resolutions (n = 86), it is regularly (n = 78) subject to some form of coercion or negative evaluation for non-compliance or failure to carry out some prescribed action. For example, in the concordance data extracted below (Figure 1), the data are sorted to display instances where IRAN co-occurs with not two places to the right of the search term.

OPEN IN VIEWER

Figure 1.

Concordance data (UNSC resolutions): Negative evaluation of Iranian actions.

While there were only two UNSC resolutions relating to North Korea between 2006 and 2012, nuclear again emerges as a collocate of the acronym ‘DPRK’ which was used to signify the state actor (UNSC 2006, n = 7; UNSC 2009, n = 7). However, rather than the phrase ‘nuclear programme’, phrases most commonly used to describe DPRK’s nuclear technology are ‘nuclear-related’ (n = 4), ‘ballistic missile-related’ (n = 4) and even (n = 1) ‘WMD-related’, as exemplified in the phrases ‘nuclear-related, ballistic missile-related and other weapons of mass destruction-related programmes’. Like Iran, the DPRK is positioned in the resolutions corpus as the receiver of the UNSC’s verbal processes, for example,

Calls upon the DPRK to join the Comprehensive Nuclear-Test-Ban Treaty (UNSC 2009);

Deploring the DPRK’s announcement of withdrawal from the NPT and its pursuit of nuclear weapons (UNSC 2009);

Demands further that the DPRK return to the Treaty on the Non-Proliferation of Nuclear Weapons and International Atomic Energy Agency (IAEA) safeguards (UNSC 2006).

In our broadsheets corpus, nuclear is also a collocate of IRAN (n = 670), which in turn co-occurs with North (n = 244) and Korea (n = 240). Nuclear appears in the phrases ‘nuclear Iran’ (n = 53), ‘nuclear-armed Iran’ (n = 25) and ‘nuclear-weapons-capable Iran’ (n = 2), existential noun phrases which never appear in the UNSC discourse. In the newspaper articles, Iran is more often associated with the phrase ‘nuclear weapons’ (n = 126) rather than the phrase ‘nuclear programme’ which is exhibited in the UNSC discourse (n = 39), as illustrated by Figure 2. Here, data are extracted from concordance data relating to the collocation of nuclear with IRAN, sorted three words to the right. It illustrates how the phrase ‘nuclear weapons’ preceded by the material processes ‘develop’, ‘achieve’, ‘get’, ‘acquire’ and ‘build’ is suggestive of a more immediate threat of attack.

OPEN IN VIEWER

Figure 2.

Concordance data (newspaper corpus): Iran associated with ‘nuclear weapons’.

What is more striking, however, is the association of Iran with North Korea that takes place in the discourse of the elite press. This stands in contrast to the UNSC discourse, which constructs each country as an individual case. Furthermore, we can extract from concordance data several examples where figurative language is used to achieve the ‘otherisation’ of the collective adversary: hyperbole, which projects exaggerated and unsubstantiated future actions onto these two states in phrases such as ‘the nuclear ambitions of North Korea and Iran …’ and ‘the more potent threats of North Korea and Iran …’; metaphor such as ‘efforts to deny gangsters in North Korea and Iran …’, ‘caught in a quagmire in North Korea and Iran …’, ‘greater threats loomed in North Korea and Iran …’, ‘keys to the deadly puzzles of North Korea and Iran …’, ‘posturing being done by North Korea and Iran …’; and even a touch of bathos, as in ‘small powers with small arsenals, i.e. North Korea and Iran …’.

As well as nation states being constituted as prominent actors within the discourse of nuclear non-proliferation, named human agents are particularly salient within the broadsheets corpus, with the two US Presidents of the period being strong keywords (OBAMA, LL 16278.6; BUSH, LL 8589.93). Our analysis suggests that the transatlantic broadsheets position the political leaders of the principal nation states as protagonists in their accounts of nuclear proliferation. However, non-nominalised signification of human agents also occurred, for example, the title PRESIDENT as well as a general category of governmental agents, OFFICIALS. Unsurprisingly, the names of the two American leaders were collocates of PRESIDENT. However, after President the next most frequently occurring collocate for both OBAMA and BUSH is the verbal process said (n = 138, n = 178, respectively), for example,

President Bush said Thursday that Iran has declared that it wants to be a nuclear power with a weapon to ‘destroy people’, including others in the Middle East;

Mr. Obama said he wanted a new United Nations sanctions resolution against Iran ‘that has bite’.

This suggests that, in the newspaper discourse which circulates within the public sphere, leadership tends to be realised in terms of an agent who speaks rather than does; a leader is constituted as the one who speaks on the matter of the day. However, political leaders are not once mentioned within the UNSC corpus. In contrast to the public sphere, prominent actors in the UNSC documents are collective, impersonal agents such as committees and different types of official. Of these, the International Atomic Energy Agency (IAEA, n = 255), also referred to as AGENCY (n = 246), are both in the top 100 keywords. In this respect, the IAEA is often referred to in the UNSC resolutions as the external authorising body tasked with providing evidence of the compliance or non-compliance of a particular nuclear actor, for example,

Reaffirms that Iran shall without further delay take the steps required by the IAEA Board of Governors in its resolution GOV/2006/14, which are essential to build confidence in the exclusively peaceful purpose of its nuclear programme and to resolve outstanding questions.

However, we also find within the 100 strongest keywords GOVERNORS and BOARD – as in the formulaic phrase ‘Board of Governors’ (above), as well as other keywords invoking agents of the ‘doxa’, for example, COMMITTEE, COMPANY, PANEL, DIRECTOR and EXPERTS. While the anonymity of this lexis serves to evacuate these agents of any personal attributes, its technocratic and occasionally hierarchical nature simultaneously imbues them with a certain power and authority through their positioning within the UNSC bureaucracy.

Discussion: Rules of formation

In our analysis, we have uncovered some of the distinctive regularities and combinations of lexis by which nuclear proliferation is constituted through a dynamic network of texts which create, maintain and transmit meanings within and across the political and public spheres. Through our comparative keyword analysis, we have identified three principal categories in each corpus: actors, strategic actions and technologies. In many cases, our qualitative consideration of collocation and concordance data has suggested that these categories were constituted differently in the different types of text. We suggest that these contradictions, changes and transformations do not indicate a lack of coherence, but rather point to certain strategies which are operationalised as these categories are delocated from one site and relocated within another site. In this respect, our approach resembles that of Mehan et al. (1990) who describe their analysis of multiple texts relating to the Cold War as ‘a constitutive approach to discourse’ (p. 158) which ‘… shows how the meaning of events is constructed in reciprocal interaction’ (p. 137). However, while Mehan et al. view this multi-voicedness as a form of ‘polyphony’ (Bakhtin, 1981; Mehan et al., 1990), we rather describe the ‘rules of formation’ of nuclear proliferation discourse in terms of ‘the conditions to which the elements of this discourse (objects, mode of statement, concepts, thematic choices) are subjected’ (Foucault, 1972: 38).

As nuclear proliferation discourse is delocated from the political sphere and relocated in the public sphere, the roles of individual actors become more prominent. In particular, the names of more nation states and their leaders feature more prominently in the broadsheets than in the resolutions. We suggest that this realises the first discursive strategy, that of personalisation. In part, it reflects the contextual positioning of the two sets of documents under scrutiny. The resolutions are generated by the UNSC members – not least the five permanent members, China, France, Russia, the United Kingdom and the United States – and speak to those in infringement of the NPT. By contrast, the broadsheets analysed describe relations between the most powerful players on the UNSC and the alleged perpetrators of the proliferation of nuclear weapons, Iran and North Korea. The difference in the occurrence of proper names reflects this relationship – the presence of the five permanent UNSC members is presupposed and therefore remains implicit within the resolutions, whereas China, France, Russia, the United Kingdom and the United States remain actors who are external to the broadsheets and therefore require explicit mention. We have also seen that the resolutions exercise an overwhelming rhetorical force, and invoke the military and economic authority of the five most powerful nations in the world against the two principal NPT infringers, Iran and the DPRK, who are both positioned as the receivers of verbal processes and recipients of repeated censure for their acts of non-compliance with UNSC statements. Since even newspapers which report on political matters aim to entertain as well as to inform (Östgaard, 1965), one outcome is to describe a state of affairs in a way which is amenable to popular consumption. An effect of this is to highlight the role of national leaders, and constitute them as protagonists in relation to the unfolding of events. This is realised not so much through ‘material’ or ‘behavioural’ processes (after Halliday and Matthiessen, 2006), but rather through ‘verbal’ processes, and in particular the verbal process ‘say’. More than anything else, a national leader is constructed in the national press as the one who speaks on the topic of the day, here nuclear proliferation. This contrasts with the elision of the names of national leaders within UNSC resolutions. The particular prominence given to the role of the IAEA also accords with insights afforded by the elite interviews from the wider CliGG database, in which the IAEA was often attributed with more authority than the UNSC itself.

Second, in relation to the realisation of strategic actions across the different text types, there appears to be a dramatisation of events for consumption in the public sphere which is not supported by the starker lexis deployed in the political sphere. The introduction of the word ‘crisis’ to describe the ‘nuclear contention’ (Rasti and Sahragard, 2012) within the broadsheets is the most striking finding in this regard. The series of elite interviews that we undertook under the wider remit of the CliGG project also intimated that neither professional nor elected officials working in the political sphere believed that nuclear proliferation was regarded as a ‘crisis’ in the political sphere. However, the broadsheets used the term not just in relation to specifically regional events that unfolded over this period in Iran and on the Korean peninsula, but also in relation to global events. In this respect, not only was nuclear proliferation constructed in the public sphere as having a ‘global’ significance, but it was also linked within the newspapers to other ‘exceptional’ phenomena such as ‘terrorism’ and ‘climate change’. In this respect, elite newspapers on both sides of the Atlantic appear to subscribe readily to the doctrine of exceptionalism (Agamben, 2005). However, on our reading, the discursive strategy of the resolution is very much the opposite, framing nuclear proliferation in normalising, technocratic language. Yet it is impossible not to be sceptical about this sanitised version of nuclear proliferation, invested as it is with both the military and rhetorical power of the five permanent members of the Security Council.

Third, descriptions of nuclear weapon technologies also appear to become reified as they move from the political sphere to the public sphere. Within the resolutions, actions relating to nuclear weapon technologies are nominalised in highly specific, complex phrases. By contrast, within the public sphere, the lexis of these phrases tends to become foreshortened and condensed, often to the phrase ‘nuclear proliferation’ – a singular, coagulated concept into which a more variegated panoply of lexis is collapsed, such as the ‘development’, ‘activities’ and ‘programmes’ which are carried out by Iran and North Korea. While this reified term detracts from the precision of the central concept of this discursive formation, its concision and repetition renders it easier to grasp as a focus of aggression and antipathy on the part of citizens and decision makers alike.

In conclusion, the findings of this article appear to support our initial thesis that a convergence of international events such as nuclear proliferation is not a static phenomenon which is constituted within any one particular set of texts. In this respect, the exploration of corpora drawn exclusively from a national press (e.g. Behnam and Zenouz, 2008; Izadi and Saghaye-Biria, 2007; Jiang, 2007; Min, 1999; Rasti and Sahragard, 2012) – while useful and insightful at a particular stage in the development of critical discourse studies – can only give a partial account of how a particular set of events is constituted in the dynamic relations that operate within and between different text types. By contrast, our exploration of nuclear proliferation has identified three discursive strategies which are deployed as meanings circulate from one site to another: in relation to actors, personalisation; in relation to strategic actions, normalisation versus exceptionalisation; and in relation to technologies, reification. These strategies are signalled by the changes, transformations and contradictions in the lexis which realise the ‘elements’ of nuclear proliferation as it is constituted within a discursive formation which is dynamic, complex and multi-faceted.