Culpability across borders: print media of the DSK affair from an SFL perspective

3.1 Interdependency and logico-semantic relations

Tables 1 and 2 summarise the incidences of the different clause complex relations of the corpus.

Hypotactic relations outnumbered paratactic ones in all articles. In his examination of scientific discourse, Martin (1996) established the correlation between hypotaxis and its capacity to convey complex meanings and communicate arguments. As hard news is a genre that seeks to identify sources of social-order disequilibrium (White 1997), inform and even persuade, this preference for grammatically complex clauses in the corpus helps achieve this, explaining the overall majority of hypotaxis. Interestingly, the extent to which it was a majority varied widely; for example, 80% of clause complex relations in Article 6 were hypotactic, compared to Articles 1, 4 and 9 which contained similar frequencies of hypotactic and paratactic relations.

The trend in Articles 1, 4 and 9 may be attested to the articles’ greater use of paratactic relations which had the effect of “balancing out” the general tendency towards hypotaxis in journalistic writing. For example, 1 and 4 employed more paratactic elaborations than other articles (with the exception of Article 7). These specify or clarify the details of an event, fulfilling one purpose of the Event Story (White 1997):

Article 9 also displayed a greater incidence of paratactic extension, which we suggest corresponds to the “justification” relationship of specification under White’s (1997) framework of how the body of a hard news story “refer[s] back to the headline/lead through a series of specifications”. In an “Issues Report” (of which Article 9 is one), justification denotes the clauses providing “evidence or reasoning which supports the newsworthy claim presented in the headline/lead nucleus” (White 1997). And so given the headline of Article 9, which orthographically “equates” Diallo with certain qualities (The Guinean victim of alleged rape by Strauss-Kahn at the Sofitel New York: “an honest, exemplary and good woman”),⁸ many of the article’s paratactic extensions correspond to this profiling. Paratactic extensions are the structural markers signalling the various justifications for Diallo’s exemplary and honest nature; construing a rape narrative around the victimised protagonist Diallo who is consistently a model citizen and thus, it would appear, incapable of fault.

In Articles 4–9, hypotactic expansions occurred the most frequently. This is mainly due to the incidence of hypotactic enhancements, which were used more than any other taxis-dependency relationship. Similarly, there was a strong tendency for extension to be combined with parataxis. Several points can be made about these observations. First, they correspond with Caffarel’s (2006) finding that in the French narrative register, enhancement tends to be hypotactic and extension paratactic (although this was also true for the American articles of the corpus). Here, the notion of language as a system of choice has guided my analysis to focus on why certain linguistic choices are selected over others. However, Nesbitt and Plum (1988, as cited in Caffarel 2006) stress the equal significance of the linguistic system as one of probabilities, where instantiations should be “interpreted against the background of expectancies”. The probabilistic nature of the clause complex system correlates with the availability of linguistic resources to realise those clause complex relations, i.e. an abundance of structural markers realising some logical relations raises its expected incidence (hypotactic enhancement) and the converse is true where there are limited options available to others (paratactic enhancement) (Caffarel 2006). So, it is not the minimal use of paratactic enhancement and hypotactic extension which is salient, but rather, that they occur instead of other relations within a framework of expectancies.

Second, the prevalence of hypotactic enhancements is unsurprising, since they express the logical causal meanings necessary to inform and persuade the readership of an ideology. It was noted that Issues Reports (Articles 2, 6, 8 and 9) used more hypotactic enhancements compared to the Event Stories of their respective cultures. Possibly, this is associated with an Issues Report’s concern to present claims rather than facts about a “supposed reality” (White 1997), where the soundness of cause-and-effect sequences within the text is crucial to the credibility and persuasiveness of that journalist’s claims/ideologies.

The final aspect of clause complex relations to be analysed here is the characteristic use of projection in journalistic writing. Since texts use language that is located on a cline from “casual talk to highly technical language” (Caffarel 2006), we can expect a news article to be positioned somewhere in the middle and exhibit elements of both written and spoken language. Typically, the spoken language of hard news articles is derived not only from direct quotes but also in the ways they address the readership.⁹ Indeed, a main source of spoken text in our corpus arises from the quotations of the participants themselves, from witnesses or even external commentators on the newsworthy event. In logical metafunctional terms, these are the locutions, and I explain below their patterns of incidence.

Overall, the American articles contained the highest proportion of locutions; 49.72% of all clause complex relations were either paratactic or hypotactic locutions. That is, they included the highest incidence of direct or indirect quotes, reflecting an American press’ tendency to practise a narrative format of reporting where quotes and details of participants are central to the construction of the news. Similarly, the French articles contained a significant proportion of locutions (38.17%), but as we will see below, this lower percentage compared to the American articles may stem from the debate format of French news, which integrates analyses and commentaries from various perspectives.¹⁰

In order to highlight how the journalistic formats of different cultures influence the depiction of an event, I look to Articles 3 (American) and 4 (French) since these both focus on DSK’s first public interview about the 2011 affair and were published one day apart. First, locutions comprise 45% of Article 3’s relations compared to 22% in Article 4. Opening with a locution taken from DSK’s interview, Article 3 foregrounds from the outset that his comments will be given prominence, forming the crux of the text. By contrast, the first locution of Article 4 appears more than a third of the way in, and neither it nor any of the other locutions are attributable to DSK himself, even though the article’s title “Mr Strauss-Kahn explains for the first time”¹¹ would suggest otherwise. Instead, the Sayers¹² are the lawyers of DSK and Diallo, such that the trial participants are commented on, rather than commenting themselves. It creates a degree of separation between the reader and the event, as if to preserve the privacy and personal dignity of the trial participants – a tenet of the Continental media’s practices¹³ – even though the interview itself transpired publicly.

The absence of plaintiff and defendant in the article is indicative of how the debate has extended beyond the initial complaint to border on wider issues, such as cross-cultural legal differences, which is referenced in clause (139): Faut-il expliquer l’état des procédures en cours aux États-Unis et en France?. Although this study did not apply an interpersonal metafunctional analysis, it is still relevant to highlight that clause (139) is one of numerous interrogatives found in Articles 4, 6, 7 and 8. The prevalence of non-polar interrogatives,¹⁴ particularly in Article 4,¹⁵ would suggest an underlying purpose to solicit perspectives and interact with the reader (debate form). In contrast, except for clause (19a),¹⁶ Article 3 (and all the American articles) are composed entirely of declaratives, demonstrating a predominant concern to inform (narrative form).

In the Guinean articles, 26.11% of clause complex relations were locutions in the Guinean articles – just over half of the incidence in the American articles. In itself, the difference does not reveal much, but drawing on White’s (1997) classification of hard news reports into the sub-types of “Event Stories” and “Issues Reports”, we may deduce that the lower percentage of locutions here is attributable to certain generic conventions. Articles 1 and 3 (American) are Event Stories since their aim is to provide a “first hand…largely unqualified, unmodulated account of what happened” (White 1997). Given the fact that the “nucleus” (White 1997) of these stories is Diallo’s phone call to her boyfriend and DSK’s television interview respectively, it becomes evident why locutions are prominent if a journalist is to replicate a past reality.

On the other hand, Articles 7–9 (Guinean) align with the “Issues Report” sub-type, which is characterised by a description of the “semiotic activity, the public controversies and debates which are triggered when some newsworthy event or state of affairs acquires the status of ‘issue’” (White 1997). Unlike in Articles 1–3, the “Issues Report” nature of Articles 7–9 does not rely on locutions to realise the issue, since they are less concerned with the reconstruction of an event than with the organisation of the differing viewpoints in the debate. While incorporating more locutions would have constituted one method of embodying the various ideas, my argument is that the multiplicity of “voices” in the Guinean corpus is manifested alternatively: there is greater emphasis on Sayers other than Diallo and DSK.

Table 3 presents a summary of all the Sayers (including implicit Sayers)¹⁷ of the verbal processes.

I identified four main categories: Diallo, DSK, legal official and other. Relative to Articles 1–6, the Guinean articles rarely placed Diallo, DSK or a legal official in the role of Sayer, whereas they were the Sayer in 93.98% of all verbal processes in Articles 1–3 (American) and 63.41% of those in Articles 4–6 (French). Instead, the Guinean articles employ participants such as newspapers or unofficial commentators (grouped as “Other”) as Sayer in 71.79% of all verbal processes. Since the 28 instances which comprise this 71.79% should not be mistaken for 28 distinct locutors (there are repetitions, such as in Article 9 where Diallo’s compatriot, “MD”, dominates as Sayer), it could be argued that there is not actually greater diversity of Sayers. The journalist may therefore still be reporting particular viewpoints to the exclusion of others, and Article 9 would appear to be evidence for this.

Nevertheless, it is significant that Diallo, DSK and legal officials are denied a capacity to speak within the Guinean corpus, rendering the issues more accessible to the general populace by avoiding a debate that is mediated exclusively by those directly implicated in the affair – the victim, accused and high-ranking officials. This also aligns with the news sites’ goal of founding a media source accessible to Africans and the African diaspora.

As the Sayer is the participant upon whom a journalist has conferred media space to “speak”, selections of Sayer may be done strategically so as to embody and endorse particular perspectives of an event. However, the choices of Sayer and Verbiage¹⁸ are often constrained by the facts of the reported event, leading me to consider that “critical analysis should pay particular attention to how what people say is transformed: there are clearly conventions for rendering speech newsworthy, for bestowing significance on it” (Fowler 1991; as cited in Chen, 2007b).

Chen’s (2005) classification of Halliday’s verbal process category into the subsets of positive, neutral and negative verbal processes illustrates how speech is “transformed” into news at the level of the verb choice. It is the first study proposing such a refinement, and the classifications nuance the ways in which a writer may reveal their attitude or purpose through the verbal process. Chen (2005) also draws a distinction between three sub-functions of negative verbal processes: material negative, attitudinal negative and relational negative verbal processes,¹⁹ and I will focus my qualitative analysis on these because they had the lowest incidence and are a good starting point in revealing how verb choices influence representations of reality.

The classification of verbal processes and their relative occurrence in the corpus is tabulated (Tables 4–6).

Across all the three cultures, negative verbal processes had the lowest incidence. These are processes which insinuate “doubt or scepticism…about the veracity” of what is said, and which we trace back to a “certain negativity of feeling on the part of the writer towards the person whose words the verbal process is being used to introduce” (Chen 2005, 30). The four instances of negative verbal processes were classified into sub-functions and are explained below.

Examples 6 and 7 from the American articles are both attitudinal negative verbal processes because they reveal the journalist’s attitude towards Diallo. A neutral verbal process like “said” could have been employed without a material change in meaning. Instead, “admitted” connotes Diallo’s reluctance or hesitance to correct her initial lie, and in this way, the journalist implies her resistance to cooperate with legal officials. In example 7, “claimed” weakens Diallo’s case against DSK in the mind of the reader by casting doubt on the truth of her statements. At the same time, “claimed” is also a relational negative verbal process here since it relates the participants through their involvement in the dispute, and Chen (2005) argues that the selection of negative verbal process over the neutral is done to dramatise the event and “capture the richness of human relationships between participants”. The low percentage of negative verbal processes across the corpus might therefore be unexpected, given the nature of the conflict at the heart of this newsworthy event. But we will see later that positive verbal processes are equally telling of the journalist’s attitudes towards the speaker whose words are being reported.

Examples 8 and 9 differ in that they are material negative verbal processes. Chen (2005) likens this category to Halliday’s material processes since they ascribe an action to a speaker. However, it is distinguished from the other sub-functions because the journalist is restricted in the choice of verb process, such that it remains unrevealing about their attitudes. The test is therefore one of agnation. Arguably, there are no other verbal processes which could have replaced “s’excuser” nor “est revenue” without materially altering the meaning. The journalist has presented the state of affairs as it stands, and we are unable to deduce any implicit viewpoints from these examples.

Overwhelmingly, the American articles preferred the use of neutral verbal processes such as “say”, “told”, “reported” and “mentioned”, with these constituting 81.81% of verbal processes. The effect of neutral verbal processes is that the reader is unable to gauge the journalist’s stance regarding what they are reporting (Chen 2005). This contributes to the sense of objectivity that is upheld as a journalistic ideal in America, since the Verbiage is being relayed as an act of human parlance without qualification on the part of the writer. Positive verbal processes, which “promote in the reader a feeling that the person whose words are being reported is wise, authoritative, benign or in some other sense positive” (Chen 2005), account for only 15.91% of verbal processes in the American corpus. Of the 14 instances which contribute to that percentage, only three are attributable to Diallo, and even then it is not Diallo’s name but the referents of “the plaintiffs” or “a New York maid”, which occupy the role of Sayer. Rather, it is various legal officials who are imputed authority through the positive verbal processes that precede their locutions. So despite the strong tendency towards neutral verbal processes, negative and positive verbal processes in the American articles still operate to reflect a journalist’s attitude towards the Sayer.

The statistics are almost diametrically opposite for the French articles, where 76% of verbal processes are positive, compared to 20% neutral. As Chen (2005) argued for negative verbal processes, we can assume here that positive verbal processes also dramatise events. They also capture the interest of the reader by raising questions about which participants are represented as authoritative, and why this is so. Even before considering these questions, Boudana’s (2010) study offers some rationalisation for this predominance of positive verbal processes in the French articles. Her interviews with 13 French war correspondents highlighted a “literary tradition of French journalism” where journalists “must demonstrate enough neutrality to prevent readers from taking offense to what is reported”, but also avoid “excessive neutrality lest readers get bored” (Boudana 2010). The journalists there themselves recognised the value in a dramatic aspect to reporting, and this is seen in our corpus through the positive verbal processes which dramatise the dynamics of the relationship between participants by establishing some kind of power disparity. In the 38 instances of positive verbal processes, there is no discernible pattern in who occupies role of Sayer. It is shared by a variety of participants including partisans of both DSK and Diallo, lawyers and even newspapers (The New York Post), suggesting that the use of such processes stems more from the aforementioned French journalistic tradition rather than a motive to uphold the authority of any one participant.

The Guinean articles contained a similar proportion of neutral and positive verbal processes (46 and 51%, respectively). I suggest that this greater balance as compared to the American and French corpuses is due to these news sites’ strategy of accounting for the expectations of a more geographically diverse readership. Not only do they write for those based in Africa as well as the African diaspora around the world, they constitute a primary source for news about the African Continent and are easily spread and retrieved online. Many of these readers may be accustomed to foreign journalistic traditions as a result of their own migration and subsequent familiarisation with overseas media cultures. On the topic of the DSK affair, it is therefore possible that the Guinean corpus incorporated elements of the American canon of objectivity (constructed in part through neutral verbal processes) and the French tendency towards interpretive journalism (positive verbal processes) in order to present a format that aligns somewhat to readers’ expectations, lest unfamiliarity with the journalistic voice leads to claims of media bias.

3.2 Transitivity and ideation

Across the corpus, material processes occurred most frequently, followed by verbal processes (Table 7). Article 3 proved the only exception, where verbal processes outnumbered material processes by one. These trends are expected of hard news articles: the use of verbal projection in this genre has already been discussed – whilst the sources quoted (Sayer) and whether they are representative of supportive and opposing perspectives are indicative of media objectivity, I demonstrated how their projected speech and the verb choice itself are revealing of cross-cultural differences in journalistic formats – and the predominance of material processes stems from its generic purpose to reconstruct and relate past and ongoing doings or happenings. Here, an ergative analysis is essential in helping to answer “Who does the journalist represent as having instigated the process?” – a grammatical expression of the media’s attribution of blame. The majority of effective clauses²⁰ in the corpus were material processes, which is predictable in the context of a crime where there is a victim (the Goal/Medium²¹ in functional terms) and a person responsible for that crime (Actor/Agent²²). Agency analysis of these material effective clauses reveals how the relationships of cause and effect between participants were portrayed.

Four categories of agents were identified: DSK and his supporters, Diallo and her supporters, legal officials or documents and the alleged rape. Others (e.g. “le New York Post”) did not occur frequently enough to establish any patterns in agency (Table 8).

Across the corpus, legal participants were more often specified as Agents of effective clauses. This is predictable from the judicial nature of the event, since DSK and Diallo are no longer the primary causers of the developments during the unfolding of the legal process; instead, it is various legal officials who have the capacity to instigate processes such as prosecution or sentencing. Furthermore, the Goal in these clauses is usually another participant of a legal nature; in this way the articles demonstrate the cause-and-effect relationship between the officials and the trajectory of the case:

In the American articles, DSK and Diallo shared similar incidences of agency (they were Agents of nine and eight clauses, respectively), but this is only 2.6% of all clauses (effective and middle) of the American corpus. DSK and Diallo were Agents of even fewer clauses across the French and Guinean corpus; however, there was a tendency to position Diallo as Agent rather than DSK. A greater number of effective clauses had DSK (agent) impacting upon Diallo (Goal/Medium) than Diallo impacting upon DSK, reflecting the fact that DSK was the accused and therefore the alleged causer of the crime. Only one of those clauses left DSK as an implicit Agent and had this been a more consistent phenomenon, especially in the French articles, it might have represented an attempt to efface DSK’s responsibility in the crime. With more empirical evidence, this potentially relates to the Socialist Party’s narrative on sexual violence – one aspect of the discourse on transatlantic differences in sexual violence attitudes which could have been emphasised by the grammatical choice of effacing the Socialist Party leader’s agency. Nevertheless, the scarce examples from the corpus render agency analysis of DSK and Diallo inconclusive. Compounded by the fact that there is no definite version of events and therefore no concrete “reality” to act as a point of comparison for what was portrayed in the articles,²³ it would perhaps be worthwhile to extend the focus of agency analysis towards determining if the DSK/Diallo cause-and-effect relationship changed significantly over the 19 months it took to reach the civil settlement. A longitudinal study of the newspapers’ coverage which concentrates on agency analysis may highlight whether culture-specific attitudes to sexual violence prevailed in attributions of causation, despite the fluctuating public opinion fuelled by revelations, conspiracy theories and the oversights at certain stages of the legal process.

Of the five occasions where the alleged rape (or other referent such as “the scandal”) was Agent, four arise from the American articles and impact on the Goal of DSK or his expected bid for presidency. Interestingly, the repercussions of the crime itself extend only to DSK and aspects of his life:

As further analysis, I compiled all breaking headlines concerning the affair from each of the chosen newspapers²⁴ to see how the event was introduced to the public in the immediacy of the accusation. As the reader’s first orientation towards the content of the article, headlines are revealing of a generalised ideology which the journalist may then detail, argue and persuade within their ensuing article. First, it is striking that in the 62 headlines, the alleged victim is never even mentioned in her capacity as “accuser” or otherwise.²⁵ Yet this is not due to lack of confirmed details – other newspapers around the world used referents such as “a hotel chambermaid” (Swaine 2011) as Diallo’s name had not been disclosed. In our newspapers, however, breaking headlines mainly referred to political and economic effects of the arrest (Table 9).

In the case of The Wall Street Journal, perhaps their emphasis on financial reporting would lead them to focus on the economic ramifications of DSK’s arrest, at least until his resignation from the IMF on the 18 May 2011. That this was also the case for the other French and American newspapers of the corpus had the effect of detracting from the seriousness of sexual violence, by censoring participants to the effect of denying a potential rape victim to be impacted and have their story legitimised as a newsworthy event in the headline. Essentially, the same news story may run in two directions; congruent to the rape narrative is the one concerning lost political opportunity and fragile economic times, and yet the former is almost invisible since there is no mention of an accuser and an arrest/une inculpation is so general and could relate to any class of crime.

Viewed from another perspective however, the aversion to headlining rape within the first two days could reflect the French and American press’ caution against contributing to a “trial by media” situation.²⁷ Each nation had an interest either to protect or to prosecute the accused, and this distinguished them from other cultures whose press was more specific in their classification of the crime (Gallagher and Willsher 2011; Sage, Naughton, and Pavia 2011).

3.3 Theme and information flow

In this section, we focus on a micro-thematic analysis of the articles to explain how Theme-Rheme structures at the clause level affirm the “correlation between genre and thematic content” (Lavid, Arús and Moratón, 2012). Further studies may apply a macro-thematic analysis, based on Danes’ (1974) theory of thematic progression, to reveal the thematic hierarchies and interrelations within a text, although previous studies only weakly support the hypothesis that thematic progression (as opposed to the thematic structure at clause level on which we focus our analysis) correlates with genre (Francis 1989, 1990; as cited in Fries 1995).

We often “define” genre in relation to its experiential content, e.g. hard news reports encompass “events or situations which are construed as threatening to damage, disrupt or rearrange the social order in either its material, political or normative guise” (White 1997, 3), and we have seen how patterns in the experiential metafunction analysis reflect the aims of news reporting genre. Here, we elaborate on the textual metafunction’s role in presenting and ordering that experience, such that when combined with the previously discussed discursive phenomena typical of French and American journalistic formats, the reader is inclined to identify the text as a hard news article.

Table 14 compares the incidence of textual and interpersonal themes in the corpus with incidences from two analogous studies on English language hard news corpora.

The prevalence of textual Themes was similar across the corpus. Their incidence ranged between 32 and 47% of clauses, which was consistent with the frequency identified in other hard news corpora (Shindo 2001; McCabe and Heilman 2007). Alongside visual methods of informational organisation such as the headline/lead nucleus (White 1997) and paragraphing, textual Themes organise the logico-semantic relations in order to “scaffold the text’s argumentative structure” (Lavid, Arús and Moratón 2012). This could explain the higher rate of textual Theme in the French articles compared to the American ones, where the former’s debate format necessitates a clearer organisation of complex ideas since there are multiple arguments interacting within the text.

Interpersonal Themes were rare across the corpus, reflecting a generic convention to exclude linguistic markers of authorial evaluation (realised through Comment Adjuncts)⁴³ and to inform rather than engaging in and soliciting dialogic exchange (realised through finites, a Wh-element or a Vocative).⁴⁴ The French language articles employed interpersonal Theme more frequently; the 11 instances were realised by a combination of Adjuncts (example 16) and Wh/Qu-elements (example 17). The former represents the journalist’s evaluation of what ensues in the clause, while the latter requests information from the addressee, creating the debate format of French journalism which has already been evidenced by an experiential metafunction analysis.

The Guinean corpus contained seven interpersonal Themes, and Article 7 displayed the highest incidence with four instances. Already atypical of hard news articles, the exclamative (302) Allons ! and the minor clause in (283) Une peccadille ! as well as the truncated polar interrogative (300) Consentie ? lead us to read the text as a kind of commentary – a genre where the journalist’s evaluations and interjections would be less anomalous. Indeed, the article’s title references the biblical narrative of David and Goliath, introducing the journalist’s own comparison of present facts to a historical event.

From the same article, the interpersonal Themes in (18) and (19) are interesting in that they resemble thematic equatives, as in Article 9’s (407) “Ce qu’elle a dû dire aux policiers est la vérité absolue !”. However, (18) and (19) were analysed as interpersonal Themes because they are agnate with adverbial groups that typically realise the Adjunct subcategory, e.g. [Évidemment]/Résultat, ils n’ont rien trouvé…, and because they summarise the journalist’s characterisation of the state of affairs.

Nevertheless, any subjectivity that the reader may associate with an interpersonal Theme is obscured by their resemblance to thematic equatives, whose structure situates the topical Theme in the experiential roles of Carrier or Token of a relational clause, thereby presenting the relation as an inalterable state of being. As a less explicit means of integrating evaluations, such a strategy adds to a sense of objectivity. Although we did not examine the incidence of modalised verbs, it is worth mentioning here that they are a source of interpersonal meaning, and that their position in the Rheme of the clause also contributes to the desired impression of objectivity.

Turning now to topical Themes, it was predicted of the news reporting genre that the majority of marked topical Themes are circumstantial elements. These temporal and spatial descriptors are accorded textual prominence in order to orient a reader who is reading an account of, rather than witnessing, the event. These marked topical Themes affix time and place in order to situate the clause in circumstantial relation to the wider context. On the other hand, some textual Themes, realised by conjunctions, indicate the temporal sequences within the text itself. Given that hard news articles are typically grounded in a single event occurring over a brief time span, they employed a combination of both – the marked topical Themes relating the event to external events (example 20), while temporally oriented textual Themes nuance the sequence of happenings that constitute the newsworthy event (example 21).

The prevalence of Sayer as a thematic head further adds to the textual distinctiveness of the journalistic genre and complements our findings on the prevalence of locutions and Sayers in the logical and experiential metafunction analyses, respectively. And so despite the semantic affinity of example (22) from the corpus and the constructed example (23), the predominant thematic structure of verbal processes across the corpus was the one seen in (22). Effectively, it prioritises the presence of the Sayer over their Verbiage, thus relegating the subjective presence of the journalist to a position of anonymity.

Several of the articles displayed a strong tendency to accord thematic prominence to the same participant. In Article 1, for example, Diallo (or a referent) was overwhelmingly the unmarked topical Theme (51 clauses) compared to DSK (1 clause). Consistently placing the alleged victim in Theme position could be interpreted as a strategy to circumscribe the involvement of DSK in the event, but we suggest it is likelier due to Article 1’s ideational grounding in “Strauss-Kahn Accuser’s Call” which necessitated reporting on Diallo’s participation in the event. We therefore shifted our analysis to focus on the interaction between experiential roles and thematic structure, and the thematic prominence of alleged victim, accused and legal participants in their capacity as Agent or Goal was examined (Table 15).

Generally, neither an impacted victim (Diallo as goal) nor a causative accused (DSK as agent) was consistently conferred thematic prominence. Perhaps this would have been too obvious an attempt to inflect these participants with unconfirmed roles in the event, thereby perpetuating the trial by media phenomenon. Across the cultures, however, an agentive legal participant was most often the thematic head, reflecting their function as overseers of the legal process.

Finally, I considered the complexity of nominal groups in Theme position. The French language Articles 4–9 displayed a greater propensity for nominally complex Themes:

This trend was particularly salient in Article 6, where lexically dense topical Themes packed together meanings which could otherwise have been realised through several clauses:

Nominalisations operate within the wider discursive strategy of grammatical metaphor⁴⁵ (Lukin, Butt, and Matthiessen 2004), and I suggest that for Article 6, they not only help to realise certain generic conventions of hard news reporting, but also inscribe the text with a generic similitude to an information report. Nominalisations in Theme position succinctly introduce the legal and media phenomenon stemming from a cross-cultural affair like the DSK case:

This confers a didactic character upon the text, helping to establish the complex argumentations which aim to inform not about the event, but on the academic discourse surrounding the event. Furthermore, heavily nominalised Themes are useful in news reporting, where media space is valuable and journalists are constrained by word limits.