1. Introduction

Who harms whom? In discourse surrounding economic sanctions, this question has no self-evident answer and that absence is not accidental. This article argues that it is the product of specific, recurring grammatical choices about agency and causation and which actors are positioned as the subjects of which kinds of processes—choices that, cumulatively, translate harm into invisibility.

Sanctions regimes typically include humanitarian exemptions: formal provisions excluding medicines, medical supplies, and basic humanitarian goods from trade restrictions. These provisions create a discursive resource that official actors routinely mobilise to deflect accountability (Nephew, 2017; Torbat, 2005). When patients cannot access chemotherapy, when hospitals exhaust essential drug supplies, when pharmaceutical companies decline orders from Iranian hospitals, official discourse can point to exemptions on paper and represent the gap between formal provision and practical effect as a problem of implementation, private-sector overcaution, or the sanctioned government's own failures. Whether healthcare harm is attributable to sanctioning-state policy, to private-sector risk aversion, or to impersonal market disruption is not a factual question resolvable by inspection: it is a discursive achievement produced and contested through the grammatical and rhetorical resources that different institutional actors deploy.

This article examines how that achievement operates and how it is contested through analysis of Human Rights Watch’s 2019 report, Maximum Pressure: US Economic Sanctions Harm Iranians’ Right to Health. The report is an analytically productive corpus precisely because it is a multi-genre document: it quotes, embeds, and recontextualises official policy documents, institutional testimony, corporate correspondence, and patient accounts within a single argumentative frame. In doing so, it makes visible, as a discursive achievement of its own, the structural tension between formal exemptions and their practical failure.

A foundational delimitation governs all claims that follow. This study does not analyse sanctions discourse comprehensively, nor does it examine primary U.S. government documents independently. It analyses how a single authoritative advocacy organisation constructs and contests sanctions discourse within its own text. Official policy language, including guidance from the Office of Foreign Assets Control (OFAC) and public statements by U.S. officials, reaches this analysis exclusively through HRW’s act of selection, quotation, and recontextualisation. That mediated access is itself analytically significant: the frequency patterns and process-type distributions reported below are findings about HRW’s editorial and argumentative choices, not about the primary corpus of official documents. Equally, following Fairclough’s (1995) insistence that critical discourse analysis be directed at all discourse, the study applies its critical instruments to HRW’s own institutional choices.

1.1 Research Questions

Three research questions organise the study. First: what grammatical mechanisms, visible in the policy-language excerpts embedded in HRW’s report, construct humanitarian exemptions as self-operating and causally disconnected from healthcare outcomes? Second: through what linguistic strategies does HRW’s counter-discourse contest those mechanisms? Third: what are the discursive limits of HRW’s own counter-discourse examined through the same analytical tools applied to the official language it contests?

1.2 Scholarly Context

The study sits at the intersection of three fields, reviewed here for the specific gap each leaves open. Within Hallidayan systemic-functional linguistics, transitivity analysis has illuminated ideological pattern-finding in political speech (Plemenitaš, 2004) and sustained literary close-reading (Kavalir, 2016), while van Leeuwen (1996) supplies the complementary account of social-actor inclusion and exclusion; this article extends the toolkit to a multi-genre advocacy corpus, where genre-internal contrast becomes itself an analytic resource. Within recontextualisation theory, Bernstein’s (1990) pedagogic-discourse origin and Fairclough’s (1995, 2003) extension to institutional discourse have been applied to intertextual embedding generally (Solin, 2004), to institutional websites (Ademilokun and Taiwo, 2023), and to political caricature within the Persian-language CDA tradition (Pourebrahim, 2017); no prior study applies the operation to a documentary genre that embeds official policy language, corporate correspondence, and testimony within a single argumentative frame. Within CDA work on advocacy and humanitarian discourse, Chouliaraki’s (2006, 2010) analysis of mediated suffering and Barnett and Finnemore’s (2004) account of NGO epistemic authority operate chiefly at the macro-institutional level; Bob (2005) and Hopgood (2006) contextualise the epistemic-authority questions examined reflexively in Section 6. This study extends these accounts to the documentary genre and supplies the close clause-level analysis their work does not undertake following Van Dijk’s (1993) insistence echoed in Wodak and Meyer’s (2015) methodological survey that critical scrutiny extends symmetrically to resistant as well as dominant discourse. Within sanctions studies, Breen’s (2021) analysis of extraterritorial compliance pressures, Emmenegger’s (2015) account of US extraterritorial financial power, and Setayesh and Mackey’s (2016) documentation of pharmaceutical access disruption in Iran establish the structural dynamics the report represents linguistically; this article does not adjudicate their causal questions but analyses the grammatical mechanisms through which causal claims about those dynamics are constructed and contested.

2. Analytical Framework

The analysis rests on two primary and mutually reinforcing instruments: Hallidayan transitivity analysis and Fairclough’s concepts of nominalisation, recontextualisation, and interdiscursivity. These operate at different and complementary levels of the same texts. Transitivity analysis provides clause-level tools for identifying which participants are positioned as agents of material processes and which are suppressed through nominalisation, passivisation, or relational constructions. Recontextualisation theory explains what happens when clauses, passages, and genres are embedded within new discursive frames, including how juxtaposition, causal connectives, and changes in process type transform the meaning of absorbed material. The two instruments are applied in sequence: transitivity analysis first establishes the grammatical architecture of a passage; recontextualisation theory then explains the argumentative transformation that embedding within HRW’s frame performs. Chilton’s (2004) deictic framework is introduced in Section 3.2 for one bounded purpose, where it captures a risk-generation mechanism that transitivity analysis alone does not fully illuminate.

2.1 Transitivity and Nominalisation

Halliday and Matthiessen’s (2014) transitivity system maps the grammatical structure of clauses onto the representation of processes in the world. The foundational distinction for this study is between material processes (representing actions performed by identifiable agents on goals) and relational processes (representing states or attributes, in which no agent acts and no goal is affected). As Plemenitaš (2004), in a foundational SFL-based text analysis, demonstrates through comparative process-type analysis of presidential speeches, the distribution of material and relational processes across a corpus encodes ideologically significant representations of agency: the aggregate pattern of process choices, not any single clause, constitutes the text's representation of who acts and who is merely present. A clause like restrictions have constrained access is material: an agent performs an action on a goal. A clause such as exemptions are in place is relational: a state of affairs simply obtains with no agent maintaining or producing it.

Nominalisation understood as the grammatical conversion of a process into a noun phrase is the mechanism that most directly enables relational constructions to suppress agency. Fairclough (2003) identifies it as a primary means, by which institutional language obscures agency and naturalises contingent arrangements as stable objects; van Leeuwen (1996) provides the complementary account of passivisation as a further agent-exclusion strategy, whereby the grammatical subject position is left unfilled rather than replaced by a noun. Both mechanisms operate in the passages analysed below. When the act of exempting, which requires political will, institutional maintenance, and active enforcement, is converted into the noun exemptions, two consequences follow: the agent of exempting disappears and the conditions under which exemptions function become invisible. The noun simply exists; it does not have conditions of operation.

Following Halliday and Matthiessen (2014), operational definitions are as follows. A material process is identified where the main verb represents action with an identifiable Agent and, typically, a Goal. A relational process is identified where the main verb represents a state of being, having, or appearing without an acting Agent. Where process type is ambiguous, particularly for semi-agentive constructions like Permit or constrain, this study applies a semantic-recoverability test consistent with Halliday and Matthiessen’s (2014) transitivity system and van Leeuwen’s (1996) account of grammatical agent-suppression: a clause is coded as material if an identifiable human or institutional actor can be recovered as semantic agent even if grammatically suppressed and as relational if no such recovery is possible. Ambiguous cases are noted in the analysis. This cumulative approach to process-type coding, attending to aggregate patterns rather than isolated instances, follows the methodological precedent established by Kavalir (2016) in her SFL analysis, where she demonstrates that the interpretive significance of transitivity choices in extended texts emerges from their recurrence and co-patterning with aspect, modality, and participant role rather than from clause-level readings in isolation.

2.2 Recontextualisation and Interdiscursivity

Fairclough's (1995, 2003) concepts of intertextuality and interdiscursivity are central to understanding the HRW report's analytical interest. The term recontextualisation originates with Bernstein (1990), who has used it in the sociology of education to describe the transformation of knowledge as it is transmitted across institutional fields; Fairclough (1995, 2003) extends the concept to the analysis of discourse in institutional and political contexts. The report absorbs official policy language, medical testimony, corporate correspondence, and patient accounts and recombines their generic conventions within a single advocacy frame. Recontextualisation understood as the transformation of material from one discursive context to another is not a neutral operation. As Solin (2004) demonstrates, when a text embeds elements of another, the receiving text actively reshapes meaning through juxtaposition, causal connectors, and changes in process type. The question this study asks of each embedded passage is: what grammatical and rhetorical operations does the embedding perform and what argumentative consequences follow? Comparable applications of Fairclough's dialectical-relational framework demonstrate that institutional texts construct legitimacy through selective recontextualisation: Ademilokun and Taiwo (2023) show, in their analysis of Nigerian university websites, that the absorption and reframing of social and professional values in institutional promotional discourse constitutes an active ideological intervention rather than neutral description. A parallel dynamic is demonstrated within the Persian-language discourse-analytical tradition: Pourebrahim (2017), in a CDA study of political caricatures, shows that pictorial metaphors operate as instruments of ideological recontextualisation, transforming political events into satirical visual arguments through the same logic of selective embedding and generic repositioning that Fairclough identifies at the verbal level. Both cases are structurally parallel to the recontextualisation operations examined here.

2.3 Corpus and Method

The primary corpus is the complete English-language HRW report Maximum Pressure: US Economic Sanctions Harm Iranians’ Right to Health (Human Rights Watch, 2019). The 55-page report synthesises interviews conducted between November 2018 and October 2019 with medical professionals, pharmaceutical industry employees, importers, lawyers, and policy experts; it incorporates corporate correspondence, bank communications, trade data, and excerpts from official policy guidance; and it contains 23 detailed patient illness narratives. All 17 policy-language quotations and all 23 patient narratives were coded, constituting the full population of each category rather than a sample.

Analysis proceeded through a two-pass close-reading and coding protocol. In the first pass, categories were developed inductively across three dimensions: process type (material, relational, mental), agency attribution (named institutional actor, named individual, nominalised abstraction, agentless passive), and causal connectors (explicit, inferential, narratively implied). In the second pass, categories were refined deductively against the theoretical framework above. The reflexive dimension, which attends to HRW’s institutional choices with the same instruments applied to official language, was built into the coding protocol from the outset: an identical coding sheet was applied to both official-language excerpts and HRW-authored passages. Because this study is single-analyst qualitative CDA, inter-subjective validity rests on transparent documentation of operational definitions and ambiguous cases, enabling readers to assess interpretive decisions.

3. Analysis I—Exemptions as Objects: Nominalisation and Relational Process in Official Policy Language

The report’s executive summary opens with a sentence whose grammatical architecture repays careful attention:

Though the US government has built exemptions for humanitarian imports into its sanction regime, broad US sanctions against Iranian banks coupled with aggressive rhetoric from US officials have drastically constrained Iran’s ability to finance such humanitarian imports.

The concessive structure, in which formal acknowledgement is immediately circumscribed by the main clause, functions as the template for the report’s entire causal argument: formal provision exists; practical effect is absent; the gap between them is the explanandum. The main clause assigns agency through the compound noun phrase broad US sanctions against Iranian banks coupled with aggressive rhetoric from US officials. The process type is material: the sanctions have ‘drastically constrained’. The agent distributes responsibility across an abstract policy instrument and its human operators so that neither bears sole causal responsibility while together they constitute the operative agent.

Compare the grammar of the embedded OFAC provision:

General licenses permit the export to Iran of certain food items, medicines, and basic medical supplies to Iran without requiring further specific authorization.

The process type requires careful analysis. Permit has a quasi-agentive meaning, but the grammatical subject, General licenses, is a nominalised administrative instrument rather than a human or institutional actor. Because no human or institutional actor grants maintains or interprets this permission as grammatical subject or recoverable semantic agent, the clause is coded as relational: a nominally agentive subject performs what is effectively a stative function. The phrase ‘without requiring further specific authorization’ reinforces this reading: the provision works automatically without institutional maintenance or interpretation.

This grammatical contrast is the nucleus of discursive contestation. In HRW’s reproduction of official language, the licence is the agent of its own operation; in HRW’s own analytical framing named human actors, specifically US officials and the policies they maintain, are the agents of constraint. The discursive contest is, at its grammatical core, a contest over who or what is the Agent of a material process: an impersonal instrument that operates automatically or an identifiable institutional actor making revocable policy decisions.

In fourteen of the seventeen policy-language excerpts reproduced in the report, the dominant process type (defined as the process type of the main clause of the primary declarative sentence) is relational rather than material. The remaining three employ material processes but with nominalised or institutional rather than human agents. Because these seventeen excerpts constitute the full population of official-language quotations in the source report, the distribution reflects HRW’s curatorial selection: the organisation has reproduced the policy-language passages that most exemplify agentless grammar because the contrast with its own material-process formulations performs the recontextualisation most forcefully.

3.1 Nominalisation as Institutional Evasion

The term exemptions is the pivotal word in the entire discursive contest. In Fairclough’s (2003) terms, it instantiates grammatical metaphor: the ongoing governmental act of exempting, which requires political will, institutional maintenance, legal interpretation, and active enforcement, is converted into a noun designating what appears to be a stable object. In fourteen of the seventeen official-language passages, the nominalised word ‘exemptions’ appears as grammatical subject of a relational process. The consistent pattern makes the nominalisation function analytically transparent: in each instance, the move from the verbal form (‘to exempt’) to the nominal (‘exemptions’) simultaneously deletes the agent of exempting and transforms an active institutional relationship into a static object.

HRW’s counter-discourse refuses this abstraction at every point where it appears. The recommendation directed at the U.S. Government reads:

Publicly clarify at the highest level that banks or companies face no legal or financial risks in exporting or financing exempted humanitarian goods to Iran.

The grammatical shift is precise and consequential. The nominalised policy object exemptions is replaced by the process construction exporting or financing exempted humanitarian goods, a gerundive retaining the ongoing, conditional character of the activity. The named agent the U.S. Government is reinstated as the subject of an obligation (clarify) rather than absent from a description of a state. The modifier publicly specifies that action is required and that its form (official public statement) matters. This is not a rephrasing; it is a grammatical recontextualisation that transforms a state description into an agency attribution and an accountability demand.

Applying Solin’s (2004) account of intertextual embedding: when the report reproduces OFAC’s relational-process language and then within the same document responds with material-process counter-language, it does not merely repeat official discourse but actively demonstrates its inadequacy through grammatical contrast. The reader is invited to perceive the difference between exemptions are in place and the U.S. Government should publicly clarify not as a difference of style but as a difference between two constructions of the world—one in which accountability is assigned and one in which it is absent.

3.2 Deictic Indeterminacy and the Production of Systemic Risk

The most analytically productive passage concerning overcompliance is a quotation attributed to U.S. Ambassador Richard Grenell:

You can do as much business as you want in Iran, but we have a say with regards to your visa. Because if you do something, we’re not going to agree to let you enter our country.

The surface grammar is formally permissive: ‘you can do as much business as you want’. The modal can assigns permission impersonally, with no named institutional granter. The second sentence introduces a countervailing material process: ‘we have a say’. The first-person plural ‘we’, attributable to the U.S. government, performs a named-agent attribution that the exemptions discourse systematically avoids. The contrast between agentless modal permission (‘you can’) and explicit agency (‘we have a say’) is, in miniature, the grammatical structure of the entire discursive contest.

The mechanism of risk-generation operates in the third clause through a deictic expression that transitivity analysis alone does not fully capture. The clause ‘if you do something’, in Chilton’s (2004) terms, deploys a proximal deixis pointing at no specified referent. The ‘something’ is not a category of prohibited conduct; it is an indeterminate possibility whose specific content is left for the auditor to imagine. An actor who cannot know in advance which transactions might constitute the ‘something’ triggering visa consequences must treat all transactions as potentially triggering. The formal permission is cancelled by the implicit warning and the cancellation operates precisely through the refusal to specify: maximum perceived risk is produced by maximum indeterminacy. This deictic mechanism transforms a nominally open compliance environment into one in which any transaction may be unsafe.

This consequence, which the report and scholars of sanctions compliance call ‘overcompliance’, is rational from the perspective of the corporate actors who practise it. They are not misunderstanding the exemptions; they are responding correctly to a risk environment in which official rhetoric has made any transaction potentially costly. Breen's (2021) analysis of extraterritorial compliance pressures establishes the structural dimensions of this dynamic; Emmenegger (2015) documents the wider pattern of US extraterritorial financial power producing de-risking behaviour in international banking that underlies it. The Grenell passage is its discursive instantiation. HRW reproduces the quotation without explicit interpretive comment, relying on the grammatical contrast between formal permission and practical risk-generation to do the argumentative work: this is an instance of implicit meaning produced through juxtaposition.

3.3 The Corporate Correspondence: Nominalisation of Refusal

The corporate correspondence that follows the Grenell passage substantiates the overcompliance dynamic in the actors’ own words. Two letters are reproduced:

JRS Pharma (June 2018): Due to political circumstances out of our control, our banks can no longer accept any payments linked to Iran business.

Gentige, Sweden (August 2018): We have been informed by our compliance dep. that we are not allowed to engage in any kind of business with Iranian-based customers.

Both letters are analytically notable for a cluster of transitivity features. Both companies represent themselves as recipients of external prohibitions rather than as agents making risk-averse choices. In the JRS Pharma letter, the grammatical subject of refusal is our banks; the company displaces agency onto its financial intermediary. The process type (‘can no longer accept’) is relational: a state of incapacity, not a decision. The causal phrase due to political circumstances out of our control employs a nominalised abstraction as the agent of the inability, framing the company’s decision as externally determined.

The Gentige letter deploys a modal of prohibition, namely are not allowed, applied to transactions that formal humanitarian exemptions explicitly permit. The phrase our compliance dep. serves as proximate causal agent, diffusing accountability within the company’s own institutional structure and suppressing the risk-calculation that underlies the decision. Both letters enact the same grammatical logic as the official exemptions discourse they respond to: agentless constructions and nominalised institutional actors replace named human decision-makers and relational states replace material choices.

HRW embeds these letters immediately after the Grenell quotation and immediately before its own analysis of overcompliance. The juxtaposition invites a reading in which the letters’ grammar of incapacity and prohibition is the institutional response to official rhetoric’s grammar of indeterminate risk. No explicit interpretive bridge is required: the grammatical rhyme between official indeterminacy and corporate prohibition performs the argumentative connection.

4. Analysis II—Constructing the Causal Chain: Recontextualisation and Counter-Attribution

HRW’s core argumentative strategy is to construct an explicit causal chain from policy decision through financial mechanism to patient outcome, a chain that official discourse systematically obscures. A representative passage from the report reads:

The restrictions on financing combined with the sharp depreciation of the Iranian currency, the rial, have resulted in severely limiting Iranian companies and hospitals from purchasing essential medicines and medical equipment from outside Iran that residents depend upon for critical medical care.

The process type is material: the restrictions (agent) have resulted in severely limiting (material process) the purchasing capacity of Iranian institutions (goal). This contrasts directly with the relational process type of the OFAC passages in Section 3. The causal syntax have resulted in is explicit and directional, constructing a one-way causal arrow from policy action to healthcare outcome.

However, the agent of this material process is itself a nominalised policy effect, the restrictions, rather than a named institutional actor. This is an analytically significant partial reproduction of the abstraction HRW contests. The contrast with OFAC passages in Section 3 is structural (material versus relational, causal versus stative) but not complete. HRW’s own causal language at this point reproduces some of the agentive suppression it criticises, most notably in passages where the word sanctions or restrictions rather than the U.S. Government serves as grammatical subject. That the recommendations section achieves fully named-agent attribution (‘the U.S. Government should’) suggests this is a discursive choice rather than a structural constraint, whether or not a deliberate one.

4.1 Absorbing and Redirecting Official Language: The Overcompliance Frame

The report’s most analytically sophisticated interdiscursive move is its treatment of the ‘overcompliance’ frame. U.S. Treasury official Elizabeth Rosenberg is quoted:

Excessive caution or ‘overcompliance’ by banks and pharmaceutical companies wary of falling afoul of US sanctions is a significant factor in limiting Iran’s access to funds for imports of medicines.

The grammatical architecture locates agency in banks and pharmaceutical companies (‘excessive caution by banks and pharmaceutical companies’) and constructs US sanctions as background condition. The nominalised abstraction excessive caution is the grammatical subject of the causative construction, making private-sector behaviour the agent of access limitation. The characterisation of this caution as ‘excessive’ implies that it is a misreading of policy rather than a rational response to it: this is paradigm causal-chain interruption.

HRW reproduces this quotation but immediately proceeds, through the Grenell passage and the corporate correspondence analysed in Sections 3.2 and 3.3, to demonstrate that what officials characterise as ‘excessive’ caution is rational response to official rhetoric that deliberately generates indeterminate risk. The argumentative move is precisely recontextualisation in Solin’s (2004) sense: the official formulation is absorbed within the report’s frame and its meaning is transformed through juxtaposition with evidence that contests its causal attribution. HRW does not state directly that Rosenberg’s account is wrong; it presents evidence whose grammar directs the reader to that conclusion. This reliance on implicit argumentation enacts the genre conventions of impartial policy analysis while advancing a contested causal claim.

4.2 Medical Specificity as Evidential Strategy

HRW’s counter-attributive discourse characteristically deploys a cluster of specificity-generating strategies: naming exact medications, citing international standards, specifying dates and quantities. The passage concerning MAHAK Pediatric Cancer Treatment and Research Center exemplifies the pattern. The hospital is documented as lacking, in May 2019, ‘pegaspargase, mercaptopurine, and vinblastine, three chemotherapy medications, all of which are on the WHO list of essential medicines’.

Three operations are performed simultaneously. First, naming exact medications prevents the vague generalisation that ‘medication shortages’ would permit: named medications with specific pharmacological profiles and procurement chains are harder to attribute to anything other than the supply disruptions their names identify. The specificity closes off the causal ambiguity that official discourse systematically maintains. Second, specifying a date (May 2019) anchors the claim to a recoverable moment, enabling independent verification and transforming an iterative access problem into a dateable evidential event. Third, grounding the claim in ‘the WHO list of essential medicines’ constructs the absent medications not as specialist items but as the minimum baseline of acceptable care: the absent medicines are internationally defined as essential so that access failure is not a matter of degree but of categorical inadequacy.

The hospital director’s quoted statement completes the argumentative movement: ‘The ones who impose sanctions say that sanctions haven’t targeted medical and humanitarian issues in Iran. Its impact is visible in the lives of children with cancer’. The first clause attributes a speech act to ‘the ones who impose sanctions’, a phrase that, by contrast with the nominalised agency of official discourse, makes the sanctioning actor a grammatical subject of imposing rather than a background institutional condition. The second clause deploys a demonstrative deictic (‘Its impact’) to link the causal chain directly from sanctioning actors to children with cancer. The transitivity chain runs:

the ones who impose sanctions (named agent) → impose sanctions (material process) → its impact (consequence) → visible in the lives of children with cancer (goal/patient). The grammar performs the causal argument that the analytical sections build discursively.

5. Analysis III—Patient Narratives as Discursive Evidence

Across the twenty-three patient narratives in the corpus, a four-phase structural pattern recurs with sufficient consistency to function as an organising template: a baseline establishing pre-disruption treatment availability, a disruption marked by sudden inability to access needed medicines, an attempted quest involving efforts to find alternatives, and non-resolution in the form of continued suffering or death. The schema is analytically legible against Labov and Waletzky's (1967) canonical model of personal narrative (orientation, complication, evaluation, resolution), from which the present structure departs precisely at the resolution phase: where canonical narrative form closes with restitution or coda, the HRW narratives systematically withhold it. Twenty of the twenty-three narratives exhibit the complete four-phase structure; three show partial restitution elements, but in each case the partial substitute is immediately marked through evaluative adjectives (‘inadequate’ or ‘insufficient’) or specification of ongoing harm as failing to restore the baseline condition. No narrative exhibits a full restitution arc. This structural consistency is a finding about HRW's construction of evidence: editorial and narrative choices systematically exclude resolutive trajectories because such trajectories would complicate the causal attribution the report advances at the macro-level.

5.1 The Epidermolysis Bullosa Narrative: Tracing the Causal Chain

The narrative concerning patients with Epidermolysis Bullosa (EB) is the corpus’s most fully articulated example of how transitivity analysis illuminates causal-chain construction. The report establishes a baseline through a relational process construction: ‘Through an NGO called EB Home, the Iranian government provided these patients with dressings for blisters free of charge’. The stable, ongoing state of affairs positions the Iranian government as the agent of material provision, a grammatical positioning that will be significant when the disruption arrives as that disruption is construed not as a failure of Iranian government provision but as a failure caused by external supply constraint.

The disruption is introduced through a temporal-causal marker: ‘After the re-imposition of sanctions, the company stopped selling to Iran a special kind of foam dressing’. The temporal conjunction ‘after’ establishes a sequence that implies but does not assert causation. The process type is material (‘the company stopped selling’) and the agent is named. This is HRW’s clearest moment of named corporate agency in the EB narrative: the company is the subject of a material process, not the recipient of an external prohibition. The grammatical positioning pre-empts the relational-state excuse: a company does not merely find itself in a state of non-selling; it stops selling.

The company’s own correspondence then makes the causal mechanism explicit: ‘This also applies to business conducted under any form of exemptions to the US economic sanctions’. The phrase ‘any form of exemptions’ performs a totalising function, explicitly enumerating the humanitarian exemptions as falling within the scope of refusal and leaving no residual category of protected transactions. This is the grammatical instantiation of the overcompliance mechanism: a relational-process clause asserting that the exemptions’ scope of coverage is overridden by the company’s compliance calculus.

The narrative closes with embodied specification of consequence. A lawyer representing EB Home testified that domestic alternatives ‘often get attached to the blisters, causing additional excruciating pain for the patients’. The causal chain runs from policy decision through company compliance response through exemption override through substitution failure to bodily pain: five links are constructed through the sequential embedding of quoted evidence. Notably, the patient’s own voice is entirely absent: the account is mediated through an NGO representative and a lawyer. This mediation is characteristic of the full corpus and is itself analytically significant as Section 6 examines.

5.2 Epilepsy Narratives: Quantification as Counter-Argument

The epilepsy narrative demonstrates a further strategy: quantification used to pre-empt official counter-arguments by constructing differential impact as numerically demonstrable. A study of 242 epilepsy patients is cited showing that 72% of those using imported anti-epilepsy drugs have reported significant access difficulties between August 2018 and February 2019 compared to 30% using domestically produced medicines. The comparative structure constructs a causal argument in grammatical form: the two groups serve as a natural-experiment comparison and the differential in access difficulties is the goal of an implied material process: the sanctions have affected one group and not the other. The causal attribution is performed by the comparative syntax itself without requiring an explicit causal connective.

The medical director of Iran’s Epilepsy Society then makes explicit what the quantitative comparison implies: ‘sanctions have caused a shortage in their medication’. The causal locution is the direct grammatical expression of the attribution that nominalised official language refuses: the agent is named (‘sanctions’), the process is material (‘have caused’), and the goal is specified (‘a shortage in their medication’). The claim proceeds to further specification, notably ‘frequent, uncontrolled seizures that risk injury and result over time in severe, permanent brain damage’, thereby performing the argumentative movement of making structural harm corporeally visible by translating the abstraction ‘sanctions have caused a shortage’ into the bodily specificity of irreversible neurological damage. The transitivity chain running from policy to medication shortage to symptom to bodily consequence constructs a causal sequence whose individual links are each analytically defensible.

6. Analysis IV—The Limits of Counter-Discourse: HRW’s Institutional Choices

The preceding sections have analysed HRW’s counter-discourse for its effectiveness in contesting official framing. This section applies the same analytical instruments to HRW’s choices as an advocacy organisation. A preliminary clarification is necessary: critical attention to HRW’s discursive strategies does not undermine the empirical accuracy of the claims the report makes. Strategic construction and empirical accuracy are analytically separable; a report can be both carefully constructed for persuasive purposes and accurate in its empirical content. The analysis below concerns the former.

6.1 Patient Voicelessness as Institutional Choice

Across the twenty-three patient narratives, a structural pattern of patient voicelessness is demonstrable through transitivity analysis. Patients experience, suffer, and die; in the narratives as constructed by HRW, they rarely speak in their own voice about their political interpretations, their attribution of causation, or their assessments of relative institutional responsibility. In the EB narrative, the patient is never the Agent of a material process. Patients are consistently positioned as the Goal of processes performed on them: pain is caused to them; dressings are provided to them; medications are withheld from them. In the epilepsy narrative, patients appear as existential or relational participants: they ‘are resistant’, they ‘have to consume’, and they ‘suffer’. The absence of patient-as-Agent in mental processes (patients are not represented as having attributions, beliefs, or interpretive positions) is as analytically significant as their absence in material-process Subject position. In each narrative, the causal attribution is provided by a medical professional, institutional representative, or legal advocate; the patient is the site of harm, not its interpreter.

This pattern has a specific grammatical consequence for HRW's argumentative structure. Were patients represented as political agents in their own accounts, attributing causation, assigning responsibility, and making contested claims, this would at minimum complicate HRW's causal attribution and might undermine its rhetorical coherence. The consistent absence of patient’s political voice reflects the genre conventions and advocacy goals of HRW's documentation practice. It is a discursive choice legible through transitivity analysis and it warrants critical notice as an institutional choice rather than transparent representation of 'what patients have said'. Chouliaraki (2010) extends her earlier account of mediated suffering to identify precisely this dynamic as a constitutive feature of post-humanitarian communication: the political self-representation of those who suffer is systematically displaced by the institutional voice that advocates on their behalf, producing a structure of representation, in which the experiencing subject is visible as a body but not as an interpreting agent. Research on institutional discourse in adjacent genres supports the generalisability of this structural dynamic beyond humanitarian reportage. Abuelwafa (2026), in a corpus-based Appraisal Theory analysis of an institutional actor's opening statement before the International Court of Justice, demonstrates that institutional speakers systematically deploy engagement resources—disclaiming contrary positions, posing rhetorical questions, and sourcing external voices—to position themselves as the authoritative interpreter of contested events and maintain argumentative dominance, a process through which the perspectives of those most directly affected surface only through the institutional speaker's own evaluative framing rather than as independent political voice. That the same logic of institutional voice over experiencing-subject voice operates across genres as different as international advocacy documentation and international legal argumentation suggests it is a constitutive feature of institutional discourse as such, not an artefact of HRW's particular genre conventions.

6.2 Narrative Selection and Epistemic Transparency

HRW does not indicate how many interviews were conducted in total and the full interview list is not published. The representativeness of the selection, including whether narratives with more ambivalent causal attributions were excluded, cannot be assessed from the published document. Presenting a selected set as a representative set without explicit acknowledgement of the selection procedure constitutes what Bob (2005) identifies as a characteristic form of epistemic closure in NGO advocacy discourse: the appearance of comprehensive documentation is maintained while the scope of documentation remains opaque. Ron, Ramos, and Rodgers (2005), in an empirical study of transnational NGO human rights reporting across a fifteen-year period, document systematic variation in what gets reported that tracks institutional positioning rather than the universe of human rights events, thereby providing empirical corroboration for the structural dynamic Bob describes theoretically.

An epistemic clarification about the own findings of this article follows: patterns in the published narratives may not be generalisable to the universe of patient experiences in sanctions-affected Iran because the published corpus reflects HRW’s selection procedure. The transitivity patterns identified in Section 5, including the absence of patient-Agent constructions, are features of HRW’s textual choices, which is precisely what makes them analytically significant as findings about advocacy discourse.

6.3 The Asymmetry of Recommendations

The section of recommendations directs substantially more detailed prescriptions at the U.S. Government than at the Iranian Government. This asymmetry reflects HRW’s causal attribution and its practical positioning, but it is also a discursive choice whose grammatical form reinforces the causal argument non-discursively. Where OFAC’s relational-process language constructs a self-operating permission (‘General licenses permit’), HRW’s recommendation calls for institutional creation: a material process requiring human agency. The U.S. Government is reinstated as the Subject of a material-process obligation; the grammar of recommendation enacts the causal attribution the report argues discursively throughout. The asymmetry of recommendations is thus not merely a political choice but a grammatical reinforcement: the side to which detailed material-process prescriptions are directed is the side discursively constructed as the causal agent.

7. Discussion and Conclusion

This analysis has traced three intersecting grammatical mechanisms, through which formal humanitarian exemptions in HRW’s representation of sanctions discourse are translated into practical invisibility. Table 1 summarises each mechanism against HRW’s corresponding counter-discursive response with the quantitative distribution of the underlying evidence.

Table 1- Official grammatical mechanisms of causal evasion and HRW’s counter-discursive responses

The through-line across all three responses evident in Table 1 is the systematic reinstatement of a named human or institutional agent in Subject position wherever official language has removed one.

The reflexive dimension of this analysis reveals that HRW’s counter-discourse is itself constituted by institutional choices that the same analytical instruments can identify. The consistent exclusion of patients from Agent positions in mental processes, a pattern demonstrated across all twenty-three narratives, removes patient’s political voice from the evidential record and subordinates patient experience to HRW’s institutional causal attribution. The selection of maximally salient cases and the opacity of the interview sample are grammatically legible features of HRW’s advocacy discourse. The point is not that these features undermine the report’s accuracy: it is that a fully critical discourse analysis of sanctions advocacy must examine all institutional voices in the discursive field, including counter-hegemonic ones, with the same analytical rigour, a principle van Dijk (1993, 252) articulates as foundational to CDA’s scope: critical analysis “should not be limited to the discourse of the powerful, but [must] also pay attention to the many forms of resistance” while applying the same standards of scrutiny to both.

These findings converge on what is, in the end, a simple and precise insight: in sanctions advocacy discourse, the struggle over causal responsibility is enacted grammatically, clause by clause, through the contest between relational and material process types, between nominalised abstractions and named agents, between indeterminate deictics and specified evidential claims. Making harm visible means, among other things, making agents grammatically visible. HRW’s report performs that visibility-work with disciplined consistency; the critical task is to apply the same grammar to the performing institution.

For CDA methodology, this study demonstrated that sustained transitivity analysis and recontextualisation theory applied with consistent operational definitions and without recourse to the supplementary macro-level frameworks surveyed in Wodak and Meyer (2015) yield analytically precise and cumulative findings when directed at a multi-genre advocacy text. The generic heterogeneity of advocacy documentation makes the contrast between process types across embedded genres particularly legible, not less so. For sanctions discourse scholarship, this analysis establishes that the discursive dimension of the formal exemption/practical effect gap is a structured, analysable phenomenon with identifiable grammatical mechanisms. For CDA work on NGOs and epistemic authority, it extends existing accounts by demonstrating that the close linguistic instruments used to analyse official language must be applied with equal rigour to the advocacy organisations that contest it.

Three directions for future research follow. Comparative corpus work juxtaposing HRW’s discourse with primary U.S. government documents would allow claims currently limited to HRW’s mediated representation to be tested against independent sources. Longitudinal analysis tracking how sanctions discourse evolves across geopolitical phases would test whether the mechanisms identified here are stable features of U.S. sanctions communication or specific to the rhetorical posture of a particular administration. Comparative analysis of advocacy discourse from organisations with different institutional alignments would test whether the grammatical contrast between relational-process and material-process language is a feature of a particular advocacy position or a structural feature of the genre more broadly.

Disclosure Statement

The authors report there are no competing interests to declare.

Declaration of AI Use

The authors used Claude Sonnet 4.6 (Anthropic) solely to improve the language and readability of the manuscript. The authors confirm that the originality and accuracy of the content have been checked and take full responsibility for the integrity of the manuscript, including the accuracy of references. The authors have reviewed the terms of use of the AI tool and confirm its suitability for scholarly publication. The methods and prompts used will be provided to the journal upon request.