Friction or Suspicion? Disfluency,Credibility and Institutional Barriers in Fraud Prevention from a Stammerer's Perspective

Abstract

Spoken interaction plays a central role in everyday social and institutional life. This article examines how interactional disfluency becomes vulnerable to misinterpretation as suspicious or risky within fraud-prevention and identity-verification contexts. While existing literature has examined the medical and psychological impacts of stammering, there is limited criminological insight into how communicative difference is operationalised within fraud-prevention practice.

Using a critical documentary and conceptual analytic approach, this article synthesises literature on stammering, disability, neurodiversity, and social harm to explore how normative assumptions about fluent, rapid, uninterrupted speech are typically embedded within institutional procedures and expectations of credibility. It identifies procedural, structural, social, and linguistic barriers experienced by people who stammer (PWS) that generate communicative friction and expose them to heightened security scrutiny during verification processes. Spoken interaction is frequently operationalised as a behavioural indicator of credibility or risk, creating significant challenges for PWS.

The article argues that communicative fluency operates as a hidden indicator of trust, rendering stammered speech structurally vulnerable to misclassification. Framed criminologically, these harms arise from routine organisational design rather than individual communicative deficit, situating speech-dependent verification as a key site where disability, risk, and legitimacy intersect.

Introduction

Conversation is a fundamental activity in human social interaction. While many take this for granted, it can be a life-altering fear for those with communication difficulties such as stammering. Acton and Hird (2004) explain that this is not necessarily due to mechanical differences in producing words, but as a result of the activities employed by those who stammer to account for the differences in speech production to align with societal and interactional norms.

Disability is not a marginal phenomenon. In the UK, approximately 24% of working-age adults identify as disabled (Department for Work and Pensions, 2024; Office of National Statistics, 2025), and between 2% and 3% of adults report having a stammer (STAMMA, 2022). These figures underscore that speech-dependent verification systems are not interacting with exceptional cases, but with a substantial segment of the population whose communicative profiles may diverge from narrow institutional norms of fluency.

This article presents a conceptual criminological analysis of how speech-dependent fraud-prevention and identity-verification systems can produce what is termed communicative harm for people who stammer (PWS). Rather than presenting new empirical data, the article adopts a critical documentary and conceptual methodology to examine how assumptions about fluent, rapid and uninterrupted speech are embedded within institutional procedures and interactional expectations. In this article, the concept of friction is used deliberately to denote both interactional disruption within spoken exchange and the procedural and institutional barriers embedded within fraud-prevention and identity-verification systems, thereby resisting any reduction of friction to speech alone. Here, communicative harm refers to exclusionary outcomes such as delayed service, failed verification, escalation, reputational suspicion, or denial of access that arise when speech-dependent systems are structured around narrow norms of fluency.

The analysis proceeds in three stages. First, it distinguishes stammering from situational disfluency in order to clarify how speech differences become vulnerable to misinterpretation within institutional contexts. Second, it examines the procedural, social and linguistic barriers embedded in fraud-prevention and identity-verification systems when fluency is treated as a proxy for legitimacy or risk. Third, it applies stammering-specific frameworks – particularly anticipation, concealment and disclosure – to demonstrate how communicative difference is managed within constrained and risk-oriented environments. Together, these stages position speech-based verification as a structurally significant site of communicative harm within contemporary criminology. The article contributes to criminological scholarship on social harm and disability by reframing speech-based verification as a site of structural exclusion rather than individual deficit. It advances the concept of fluency as an implicit credibility test embedded in contemporary risk governance, extending criminological analysis into the domain of speech-dependent verification systems.

Historically, criminological research has tended to frame disability as an individual health concern, often obscuring the wider social and structural inequalities that shape everyday institutional interactions (Macdonald & Peacock, 2025). Drawing on Hillyard and Tombs’ model of social harm (2007), Macdonald and Peacock (2025) argue that disabled individuals disproportionately face financial, cultural, physical, sexual and psychological harms. Crucially, such harms are not only embodied but produced through routine organisational encounters in which barriers to access, credibility and participation become normalised. The linguistic and communicative dimensions of discrimination remain comparatively under-examined. As Caldwell et al. (2024, p. 369) note, communication disability is frequently excluded from conversations about diversity, equity and inclusion, reflecting its limited engagement as a socially constructed phenomenon within criminological scholarship.

Clarifying stammering and disfluency in institutional contexts

It is important to distinguish stammering from disfluency more broadly, in institutional contexts where speech patterns are interpreted as indicators of credibility or risk. Disfluency is a universal feature of speech and increases among non-stammering speakers in high-pressure, high-stakes, or deceptive situations, where hesitations, self-repairs, pauses and repetitions become more frequent. These forms of situational disfluency are often culturally associated with nervousness, uncertainty, or dishonesty and play a significant role in shaping listener expectations and stereotypes. Stammering, by contrast, is a neurodevelopmental communication difference characterised by involuntary disruptions to speech flow, including repetitions, prolongations and blocks that persist across contexts and are not contingent on deceptive intent or situational stress alone. While stress or pressure may exacerbate stammering momentarily, these disruptions are not strategic, adaptive, or situational in the way that disfluencies produced by typically fluent speakers often are. Stereotypes linking disfluency with dishonesty are often grounded in observations of fluent speech under deceptive or evaluative conditions. When fluent speakers hesitate, pause, or self-correct during lying or high cognitive load, listeners may learn to treat these behaviours as diagnostic cues of insincerity.

Jeske et al. (2018) explore the anxiety and performance of interviewees in telephone job interviews, finding that vocal fluency was used to gain an impression of candidates and disfluency, often seen in PWS such as hesitations, incoherent sounds, repetitions, filler words and unfinished sentences, negatively impacting the interviewer's perception of their abilities. Bridging between contexts, Amasiatu and Shah (2019) explain how investigators have been trained to ‘look at behavioural cues such as signs of nervousness, stammering and contradictory statements, to determine if the customer is lying’ (2019, p. 19) which led to a significant reduction in first party fraud losses (Amasiatu & Shah, 2019), but the impact on genuine customers whose normal communications are pathologised as indicators of fraudulent behaviours, is unknown.

These learned associations are then incorrectly generalised to stammered speech, despite its distinct aetiology and interactional function. As a result, PWS may be assessed through a behavioural framework that is not typically designed to accommodate neurodivergent speech patterns. Speech features intrinsic to stammering may be misread through a lens shaped by assumptions about situational disfluency, leading to heightened scrutiny, misclassification, or escalation. This process can be understood as reflecting the institutionalisation of interpretive norms that treat fluency as a proxy for sincerity, competence and legitimacy within contemporary risk-oriented governance.

The present analysis does not claim that fraud-prevention systems identify stammering as deceptive behaviour, nor that all disfluent speech is penalised. Rather, it proposes that the convergence of behavioural fraud markers, listener heuristics and speech-dependent verification processes creates conditions under which stammered speech is vulnerable to misinterpretation. The argument advanced here is not that disfluency itself is inherently problematic, but that institutional systems trained – formally or informally – to treat disfluency as a risk cue may inadvertently criminalise a specific neurodivergent mode of communication. Having established how stammered disfluency becomes vulnerable to institutional misreading, the following section situates stammering within neurodiversity and disability frameworks to clarify where the locus of harm is analytically located.

Stammering as neurodivergence

Positioning stammering within a neurodiversity framework reframes it as a form of naturally occurring neurocognitive variation rather than a singularly pathological deficit. Within the broader neurodiversity literature, neurodiversity is defined as ‘the biological reality of infinite variation in human neurocognition’ (Singer, 1999, as cited in Isacoff et al., 2026), encompassing differences that may involve vulnerability as well as potential strengths, depending on environmental and social context (Isacoff et al., 2026). Importantly, neurodivergent traits are not conceptualised as inherently adaptive or maladaptive, but as acquiring meaning through their interaction with cultural norms, institutional expectations and material conditions (Isacoff et al., 2026). This understanding aligns with contemporary disability theory, particularly the social and relational models of disability, which locate disability in the interaction between individual characteristics and socially constructed norms rather than within impairment alone (Oliver, 1990; Shakespeare, 2014).

Within stammering research, there is growing recognition that the lived impact of stammering cannot be explained solely by speech disruption, but is mediated by stigma, listener reactions, communicative expectations and social power relations (Boyle, 2015; Yaruss & Quesal, 2004). These insights parallel neurodiversity-informed approaches, which emphasise that communicative difference becomes disabling primarily in environments that privilege narrow norms of fluency, timing and verbal performance. Adopting a neurodiversity framework does not entail the rejection of intervention, but instead reshapes its ethical justification. Neurodiversity-oriented scholarship emphasises autonomy, participatory decision-making and respect for diverse goals, rather than the presumption that normative functioning should be restored wherever possible. Vanryckeghem (2026) aligns with this position, arguing that people who stammer must be understood as heterogeneous in their experiences and aspirations, citing Bortz (2024) who notes that ‘neurodiversity allows for differences among people who stammer in terms of those who perceive their difference as a challenging and lonely experience and those who want to be fluent speaker’ (Bortz, 2024, p. 519). Reframing stammering as neurodivergence does not negate the reality of distress or functional impairment; rather, it foregrounds the ethical necessity of ensuring that intervention targets are collaboratively defined, proportionate and responsive to individual meaning-making rather than imposed normative expectations. Instead, stammering is situated within human neurodiversity to challenge reductive deficit-only models while preserving the legitimacy of evidence-based intervention grounded in client-defined goals. While this theoretical reframing relocates stammering within neurodiversity, it remains necessary to examine how broader social norms and cultural narratives shape its lived and institutional reception.

Situating stammering in societal contexts

Acton and Hird (2004) explore the inconsistencies: stammering is interrupted speech flow; however, there is no standard for ‘normal’ or ‘fluent’ speech and this cannot necessarily be used as an endpoint on a single two-dimensional scale. Given that, additionally, ‘everyday speech does not evince perfect fluency’ (Acton & Hird, 2004, p. 500), stammering becomes marked as abnormal not in relation to everyday speech practice, but in relation to an idealised and socially enforced norm of fluency. The labelling of stammering specifically as a speech disorder or form of dysfluency, therefore, reflects normative judgements about how speech should sound, rather than empirical deviation from ordinary communicative behaviour. Therefore, stammering has been cemented in society as a condition that needs to be treated or ‘cured’, a source of comedy and a characteristic of anxiety or deception. This has often been reflected in popular culture and the media, through various characters who are ‘depicted as socially, mentally, or morally flawed’ (Seitz & Choo, 2022, p. 6). For example, Porky Pig (Looney Tunes), Ken Pile (A Fish Called Wanda), Albert Arkwright (Open All Hours), Jim Trott (The Vicar of Dibley) and Professor Quirrell (Harry Potter and the Philosopher's Stone). While anxiety does not cause the onset of stammering, it may exacerbate (National Institute on Deafness and Other Communication Disorders [NIDCD], n.d.; Neef & Chang, 2024; Smith & Weber, 2017). Research consistently demonstrates higher rates of anxiety and social anxiety among adults who stammer. These elevated anxiety levels are widely understood in the literature as consequences of stigma, repeated negative interactional experiences and structural barriers, rather than intrinsic features of stammering itself. This distinction is important, as it challenges the misconception that anxiety is an inherent or causal component of stammering. It is central to understanding how institutional actors may conflate disfluency with suspicious behaviour. In high-stakes institutional contexts such as telephone-based verification, anxiety may exacerbate speech disruption, increasing the likelihood of misinterpretation.

Public perceptions are frequently shaped by narrow and stereotypical portrayals, leading to ignorance and generalisations rooted in comparisons with dominant norms of fluent speech and cultural representations such as these contribute to misconceptions and negative attitudes toward stammering. Such attitudes often frame stammering as negative (Arnold & Li, 2016; De Nardo et al., 2023; Mancinelli, 2019) and a deviation from the expected standard of communication, reinforcing social exclusion. As noted by the Alliance for Financial Inclusion (Raman, 2024), ‘stigma perpetuates exclusion’, a dynamic that continues to marginalise PWS in both social and institutional contexts:

“listeners construct an image of people who stammer which becomes for them a matter of ‘fact’. So situation avoidance becomes an indication of poor social skills, word substitution a confirmation of guardedness or reticence and a tendency to keep responses short provides evidence of shyness or introversion”.

(Acton & Hird, 2004, p. 506)

As demonstrated above, techniques adopted by PWS to attempt to be ‘fluent’ can be misconstrued as further non-compliance with other social norms, such as expected patterns of verbal confidence, emotional expressiveness, or interpersonal engagement. These socially embedded misinterpretations do not remain confined to everyday interaction; they acquire particular significance in institutional environments structured around risk detection and credibility assessment.

Perception of stammering in criminological contexts

In fraud prevention and security verification contexts, this misalignment between communicative norms and disfluent speech can have material consequences: access denial, increased scrutiny, or misclassification as high-risk customers. Telephone-based identity verification processes in banking typically follow a standardised sequence involving automated prompts, live agent interaction and knowledge-based authentication. At each stage, normative expectations of fluent, immediate and uninterrupted speech can become embedded into the structure of the interaction (Muskett et al., 2013). Automated Interactive Voice Response (IVR) systems often rely on speech recognition software and impose time-limited prompts, offering little accommodation for disfluent or delayed speech. This can result in the call being ended prematurely or the caller being routed incorrectly. Once connected to an agent, customers are usually required to provide personal details without the opportunity to preface responses with a disclosure of communication difference or request for adjustment. During knowledge-based authentication, callers are asked to recall and recite factual information under conditions of institutional authority and time pressure.

Interactions that are shaped by rigid expectations of verbal fluency, immediacy and clarity, with limited tolerance for non-normative speech patterns, often generate barriers for individuals who stammer. This dynamic is particularly pronounced in high-stakes, protocol-driven encounters such as security and identity verification procedures in financial services. Individuals who deviate from these expectations, whether through repetition, pauses, or changes in vocal rhythm, may be miscategorised as suspicious or untrustworthy through their attempts to comply. Within systems designed around fluency as a default communicative baseline, stammering risks being pathologised; treated as a sign of deceit or non-compliance. For PWS, these conditions can intensify disfluency. Repetitions, hesitations, or speech blocks may be misinterpreted as signs of nervousness, evasion, or deception, rather than recognised as features of a communication difference. Such processes illustrate how stammering, when encountered in systems designed around conventional speech, may be marginalised and pathologised, shifting the burden of adaptation onto the speaker and positioning fluency as a prerequisite for full participation in accessing financial services. To examine how these procedural expectations are structured and operationalised, the following section outlines the documentary and conceptual methodology adopted in this study.

Methodology

Given the restrictions on financial institutions disclosing the specific content of their security procedures, this study draws upon publicly available sources to examine the challenges encountered by PWS in telephone-based banking interactions. Adopting a critical documentary and conceptual analysis rather than an empirical or mixed-methods design, the analysis incorporates insights from academic literature, UK Government publications, reports from the Financial Ombudsman Service, media articles, materials produced by the UK stammering charity STAMMA, commentary from the financial news outlet FINEXTRA and user accounts shared via online forums such as Reddit. These sources are treated not as empirical data, but as documentary sites through which institutional norms, assumptions and power relations can be examined, consistent with critical criminological and disability-studies traditions (Hillyard & Tombs, 2004; Pemberton, 2015). This information is utilised to map out typical verification procedures, identify common security questions and examine how the standard information banks request – such as dates, names and account details – and the manner in which these requests are operationalised may create challenges for PWS. Rather than aiming for descriptive cataloguing alone, the analytic objective is theoretical interpretation: specifically, to identify how communicative expectations embedded within fraud-prevention and identity-verification systems may disadvantage speakers whose communication diverges from conventional fluency.

Source selection focused on materials that either (a) reflect formal institutional protocols or (b) provide authentic user perspectives on security encounters. By triangulating between institutional documentation, academic literature and lived experience accounts, the approach aims to offer a holistic picture of systemic barriers. Themes converged across sources – such as time-pressured speech, fixed authentication responses, or suspicion arising from hesitation – these were treated as analytically significant.

Drawing on the criminology of social harm (Hillyard & Tombs, 2004; Pemberton, 2015), this article positions the misinterpretation of stammered speech within a framework of structural victimisation. Rather than locating harm within individual interactions, this perspective emphasises how institutional architectures and communicative norms reproduce inequality. In line with neurodiversity-informed criminological approaches, systems that rely on speech-based verification, risk assessment, or security screening are implicitly designed around neurotypical communicative profiles, which renders those who deviate from these norms vulnerable to suspicion, exclusion, or punitive treatment. These exclusionary effects constitute forms of communicative harm, where organisational and technological infrastructures produce social harm through routine procedural design rather than through intentional discrimination.

Although this methodology provides rich insights, it is important to note its limitations: access to proprietary fraud detection and verification protocols is restricted and findings therefore rely on inference from public-facing documentation and secondary reporting. Accordingly, the analysis does not claim to reconstruct or observe fraud-prevention interactions in situ, nor does it evaluate the internal functioning of proprietary detection systems or internal decision-making processes. Claims regarding misclassification, suspicion, or exclusion are therefore framed as theoretically informed propositions rather than empirical observations of institutional practice.

The analysis extends criminological inquiry beyond traditional notions of crime to include the ways in which ableist procedural design can criminalise difference and reinforce systemic inequities, while often remaining unacknowledged within regulatory and ethical frameworks. By situating speech-based verification as a site where disability, risk and legitimacy intersect, the article contributes a framework for understanding communicative harm as a structurally produced phenomenon rather than an individual communicative failing. Addressing the former first, security systems, verification protocols and customer interaction procedures frequently operate on the assumption of fluent, uninterrupted speech as a marker of legitimacy and competence. For PWS, such procedural expectations can result in repeated misrecognition, delays and exclusion from essential services. The following section examines these procedural barriers in greater depth, illustrating how institutional processes designed to manage risk instead reproduce it by embedding structural disadvantage within their communicative foundations. Each section illustrates how communicative norms are operationalised within institutional frameworks.

Analysis

Procedural and structural barriers in speech-based verification

Many institutional processes, in particular identity authentication, are typically designed around assumptions of fluency, rapid response and uninterrupted speech, leaving little accommodation for communicative diversity. In terms of ongoing monitoring controls, organisations will often require a customer to communicate telephonically and answer security questions to prove their identity prior to proceeding with the call. These will likely fall into the category of personal information: full name, date of birth, contact details, account number and so on. For many PWS, these high-stakes communicative contexts can be challenging. While the expectation is immediate responses from genuine customers, adaptive strategies such as word substitution or circumlocution, which help PWS maintain fluency in everyday conversation, cannot be used effectively in this setting because the answers demanded are fixed and non-negotiable. Research on stammering anticipation refers to the expectation that a stammer may occur on particular sounds, words, or in particular contexts and is widely documented as a central feature of adult stammering experience (Jackson et al., 2015). In institutional encounters involving fixed identity questions, time pressure, or heightened scrutiny, anticipation can shape not only how speech is produced but how interaction itself is managed.

This procedural inflexibility represents a significant barrier to participation for individuals with communication disabilities (Bérubé-Lalancette et al., 2025), excluding individuals whose speech does not conform to these expectations, which can lead to avoidance (Perez et al., 2015). The restrictive and time-sensitive nature of these exchanges heightens the likelihood of stammering, pauses, or blocks, which may be misinterpreted as evasiveness or deceit (Amasiatu & Shah, 2019). Within fraud-prevention and identity-verification contexts, where speech may be implicitly relied upon as a behavioural indicator of legitimacy or risk, PWS may therefore struggle to satisfy procedural expectations and risk being flagged as suspicious or non-compliant by automated systems or call agents.

A notable document stating the overlap between stammering, financial inclusivity and criminological contexts is a guide by STAMMA to educate those in customer-facing roles and how telephone calls may impact those who stammer. This guide highlights potential problem areas such as artificial intelligence and fraud detection systems, which may misinterpret pauses and vocal stammering, as well as customer identification procedures solely requesting information that cannot be changed – such as personal information (STAMMA, 2023).

Advocacy organisations such as STAMMA (2023) propose a range of practical reforms to address this. These include customer profile flags or tags to alert staff to speech differences, alternative authentication procedures and the provision of non-verbal communication channels. Technologies such as RelayUK (British Telecommunications, n.d.) enable users with various communication differences to interact with organisations through mediated digital text or translation services, reducing the risk that a speech difference is misinterpreted as an indicator of fraud or risk. Such measures align with the Financial Conduct Authority's (FCA, 2023) Consumer Duty of Care and broader fair treatment of customers' obligations, which require firms to ensure equitable access and outcomes for all consumers, particularly those identified as vulnerable (FCA, 2006). The FCA (2021) further emphasises the need to improve understanding of outcomes for consumers in vulnerable circumstances, underscoring that barriers of communication or access can amount to breaches of fair treatment principles. In addition to regulatory expectations under the FCA's guidance, financial institutions are also subject to the Equality Act 2010, which imposes a statutory duty to prevent discrimination and make reasonable adjustments to ensure disabled people are not placed at a substantial disadvantage when accessing services, including where communication practices or verification procedures indirectly exclude people who stammer.

This procedural landscape reveals a critical asymmetry: fluency is treated as the default standard and deviation is operationalised as risk. By embedding these communicative expectations within verification scripts, IVR systems and risk models, institutions risk converting a benign speech difference into a marker of suspicion.

Musilli (2022) describes concealment as both protective and destabilising, noting that opportunities to conceal can feel like ‘catnip’ despite generating feelings of inauthenticity. Yet in fraud-prevention contexts, concealment strategies are constrained by fixed-response requirements, intensifying the tension between fluency management and institutional expectation. Yet concealment becomes fraught in fraud-prevention contexts, where PWS may use word substitution, circumlocution, or avoidance to reduce the likelihood of stammering and manage stigma, preserve social identity, and avoid negative listener reactions (Mancinelli, 2019; Musilli, 2022). However, because responses are often constrained, fixed, and non-negotiable, the requirement to repeat exact names, dates, or account details removes the flexibility PWS commonly use to maintain fluency. Paradoxically, these concealment strategies-such as delaying responses or rephrasing earlier in an interaction-may themselves be read as suspicious behaviour, reinforcing the very stereotypes PWS seek to avoid.

Assistive technologies and the reproduction of bias

While advances in assistive technology have the potential to improve accessibility, they can also reproduce underlying biases. Lea et al. (2023) note that many speech-recognition systems are trained primarily on fluent speech data and, therefore, disfluent or atypical speech patterns may be inaccurately categorised as errors or failures. In such cases, communicative breakdowns are implicitly attributed to the speaker rather than to the technological system or its design assumptions, reinforcing perceptions of personal inadequacy and unreliability. Technologies designed to enhance access may therefore inadvertently amplify exclusion when they reproduce neurotypical communicative norms. Through a criminological lens, these sites of assistance can be understood as mechanisms of procedural discrimination that enact structural harm by embedding ableist assumptions about how speech should occur. Technologies such as speech-recognition systems or rigid IVR time-outs seek to filter out non-normative speech rather than adjust interactional pacing or evaluative criteria, effectively minimising the manifestation of stammering rather than redesigning systems to accommodate it.

Disability models, ableism and structural harm

Contrary to the medical or individual model of disability, owing to the fluidity of the social model, it is possible to ‘theorise a possible future with greatly reduced disability or without disability’ (Bailey et al., 2015, p. 14). Both disability and vulnerability can provoke discomfort or avoidance in organisational contexts, often due to the persistence of negative societal attitudes and ableism (Kversøy & Kversøy, 2025). Bailey et al. (2015) suggest a key reason for this lies in negative experiences of disability, damaging media representations and generalised ableism in society, noting that disability can be experienced as variable rather than fixed: ‘I am disabled, sometimes a lot and sometimes a little’ (Bailey et al., 2015, p. 14).

This highlights that procedural barriers do not exist in isolation but are deeply entwined with social expectations about fluency and communicative competence, which shape how stammering is interpreted in interaction. It may therefore be concluded that the pathologisation of PWS is not inherent to stammering itself but emerges as a by-product of ableist and disablist constructs embedded in institutional design. The procedural infrastructure of fraud prevention is not neutral; it actively shapes communicative interaction and creates conditions under which stammering is structurally vulnerable to misclassification.

Social and linguistic barriers in institutional interaction

Social and linguistic barriers operate at the level of interaction, shaping how speech is heard, interpreted and evaluated within institutional encounters. These barriers are not reducible to individual prejudice but emerge from widely shared, often unspoken assumptions about what constitutes appropriate, credible and cooperative communication, assumptions that acquire particular significance in fraud-prevention contexts where speech is used as a proxy for legitimacy.

Social interaction is shaped by conventional expectations of fluency, immediacy and coherence, guided by Grice's (1975) conversational maxims of quantity, quality, relevance and manner and bound by interactional preference. Deviations from these expectations, such as extended pauses, repetitions, or reformulations common in stammering, can be interpreted by listeners as breaches of conversational norms and associated with deceit, uncertainty, or lack of sincerity (Ziano & Wang, 2021), or used to form judgements about intelligence, honesty, or professionalism (Caldwell et al., 2024). Rather than treating Grice's maxims as a prescriptive framework, this analysis uses them heuristically to illustrate how deviations from expected timing or delivery may be perceived as communicative breaches, even where such features reflect a speech difference rather than conversational intent. Linguistically within Conversation Analysis, ‘preference’ refers not to personal inclinations but to the organisation of interaction around socially conventional participation. For example, refusals or disagreements are typically delayed or mitigated because they carry greater potential for social tension (Heritage, 1984). When viewed through this lens, stammering becomes a site of tension between individual expression and conversational expectation rather than a deficit in communicative intent.

PWS face social and interactional pressure to conform to these preference structures by responding promptly, fluently and in alignment with normative timing. Deviations may be interpreted as indicators of rejection or disagreement, contradicting the principle that the responsibility should rest ‘on the organisation to assess the environment, policies and processes to ensure fair and equitable treatment’ rather than on the individual (Caldwell et al., 2024, p. 373). In institutional settings where risk detection is influenced by perceived behavioural normality, these interpretations can shape material outcomes, including escalation or denial of service. This reflects what might be understood as a ‘criminology of credibility’, in which communicative norms function as informal mechanisms of social regulation, determining who is heard as trustworthy and who is not.

Telephone communication and heightened vulnerability

Research consistently identifies telephone communication as a significant barrier for people with speech impediments due to the absence of visual cues such as gesture and facial expression, which can amplify misunderstanding and interactional strain (Constantino et al., 2022; Crichton-Smith, 2002; Rosenberg & Kohn, 2016). Disfluency may be misinterpreted as deception, disengagement, technical failure or call abandonment. Telephone communication therefore amplifies existing social and linguistic barriers. When visual and contextual cues such as facial expression, gesture and embodied timing are absent, listeners rely more heavily on speech fluency and rhythm to interpret intent and engagement (Trujillo et al., 2022). Silences or pauses may be misattributed to technical failure, hesitation, or disengagement, increasing the likelihood of interactional breakdown or premature call termination (Petrunik, 1982). For PWS, this environment intensifies the risk that stammered speech will be misinterpreted not only interpersonally but also institutionally, through responses such as escalation, repeated verification, or denial of service.

This dynamic operates across both micro-level interaction and macro-level institutional response. PWS are placed in a structurally disadvantaged position not because they cannot communicate, but because their communicative style is judged against a narrow norm. Research into courtroom testimony by PWS (Constantino et al., 2022; Maras et al., 2020; Maras et al., 2025) demonstrates how systems that prioritise fluent delivery risk conflating disfluency with credibility. The emphasis on procedural flexibility and accommodation within the criminal justice system underscores the need for systemic, rather than individualised, responsibility in addressing communicative bias. These insights support the argument that procedural barriers in speech-reliant verification systems constitute a form of structural harm that inadvertently risks criminalising communicative difference. Here, ‘criminalisation’ does not imply that stammering is treated as criminal behaviour per se, but that disfluency becomes operationalised within risk-oriented systems that associate deviation from communicative norms with suspicion, escalation and exclusion.

Disclosure, legitimacy and institutional responsibility

Disclosure of a stammer is sometimes used by PWS as a strategy to mitigate stigma and manage listener expectations (De Nardo et al., 2023; Mancinelli, 2019). Research suggests disclosure can sometimes improve listener perceptions and reduce negative judgement, in supportive or informed environments (e.g. Boyle & Gabel, 2020; Mancinelli, 2019). While disclosure may pre-empt misjudgement in receptive contexts, individuals should never be compelled to disclose medical information to access essential services. Where institutional norms privilege fluency, disclosure alone cannot overcome structural bias, particularly when organisational responsibility for accessibility is displaced onto individuals. These pressures may encourage strategies such as disclosure, avoidance, or concealment, each of which carries interactional risks and costs and further illustrates that communicative difficulty is produced not by stammering itself but by the evaluative frameworks through which speech is judged. Within tightly scripted or technologically mediated interactions, disclosure may have limited practical effect: where systems prioritise speed, standardisation, or behavioural consistency, advance disclosure cannot fully counteract structural constraints.

Fluency functions as an unspoken criterion of legitimacy, positioning stammering as deviation from the social ideal of competence. The analysis does not claim that all listeners interpret stammering negatively, nor that institutions penalise disfluent speech; rather, it argues that when institutional systems rely on speech-based cues as indicators of normality or risk, communicative differences become structurally vulnerable to misclassification through the convergence of interactional norms, institutional expectations and technological mediation. Anticipation, concealment and disclosure can thus be understood as interactional risk-management strategies employed by PWS in environments where communicative deviation carries potential consequences; these adaptations reveal how individuals negotiate systems that implicitly treat fluency as a marker of credibility, yet may become further exposed to misclassification when interpreted through fraud-prevention logics insensitive to neurodivergent communication. Communicative harm does not arise from stammering itself, but from institutional arrangements that conflate normative speech patterns with trustworthiness, compelling individuals to manage, conceal, or explain speech in ways that can paradoxically heighten suspicion.

What organisations can do

Addressing communicative harm requires a shift from individualised accommodation towards systemic responsibility in the design of fraud-prevention and identity-verification processes. Organisations should critically examine where assumptions of fluent, rapid and uninterrupted speech are embedded within verification scripts, time-limited prompts, escalation thresholds, behavioural monitoring systems and agent training. If fluency functions as an implicit credibility test within risk-oriented governance, then redesigning speech-dependent systems is not a matter of optional adjustment but of recalibrating institutional indicators of legitimacy. Decoupling speech fluency from assessments of credibility or risk is therefore essential, particularly in high-stakes interactions where disfluency may be misinterpreted as evasion or deception.

From a criminological perspective, such reform entails scrutinising the informal credibility heuristics that operate within fraud detection frameworks. Where hesitation, repetition or delayed response are implicitly treated as behavioural risk markers, institutional systems may inadvertently penalise neurodivergent communication. Inclusive redesign should thus be understood as a form of harm reduction within security infrastructures, ensuring that difference is not operationalised as suspicion.

Practical measures may include offering alternative authentication pathways that do not rely solely on speech (such as secure app-based verification, asynchronous written confirmation, or mediated communication services), allowing additional response time without penalty, enabling advance communication preferences within customer profiles, and training frontline staff to recognise communicative diversity without triggering suspicion. Organisations might also review automated and human-led escalation criteria to ensure that disfluency alone does not function as a proxy indicator of deception or non-compliance.

Crucially, these adjustments should not depend on disclosure or self-advocacy by people who stammer. Where access to essential financial services is contingent upon communicative performance, responsibility must rest with institutions to design processes that anticipate communicative diversity as a routine feature of social life. By embedding inclusivity as a default characteristic of verification systems, organisations can reduce exclusion, limit unjust escalation and strengthen the ethical legitimacy of fraud-prevention practice.

Conclusion

This article has shown that the harms experienced by people who stammer (PWS) in fraud-prevention and identity-verification contexts are not incidental but structurally produced through institutional reliance on fluent, rapid and uninterrupted speech as a marker of legitimacy. When communicative norms are typically embedded within procedural scripts, technological systems and frontline decision-making, disfluency becomes vulnerable to misinterpretation as uncertainty, evasion, or risk. These outcomes reflect not individual communicative failure, but the consequences of organisational designs that treat fluency as a prerequisite for credibility. Such systems risk converting a neutral difference in speech into a basis for suspicion and exclusion.

Importantly, the analysis demonstrates that mitigating communicative harm cannot rest on individual disclosure, adaptation, or resilience. While PWS employ strategies such as anticipation management, concealment, or disclosure to navigate interactional risk, these practices arise in response to structurally constrained environments rather than personal preference. Responsibility therefore lies with institutions to critically reassess how speech is operationalised within risk assessment and authentication processes and to recognise communicative diversity as a routine feature of interaction rather than an exception requiring special justification. Placing the burden on individuals not only perpetuates harm, but obscures the role of institutional power in shaping whose speech is deemed credible. By reframing fluency-based assessment as a site of organisational responsibility, this article positions inclusive procedural design as a form of harm reduction within criminological practice. Decoupling communicative differences from suspicion does not weaken fraud prevention; it strengthens its ethical legitimacy and reduces unjust exclusion. Framing stammering through a neurodiversity lens exposes how barriers arise not from the speaker but from systems designed around normative communication. This perspective aligns with the social model of disability, which situates disadvantage within the environment and social structures rather than the individual. It also resonates with contemporary neurodiversity scholarship in criminal justice, emphasising that structural inequities, exclusion and the pathologisation of difference produce real-world harms for neurodivergent people.

Future research should prioritise participatory and empirical approaches – including interviews with PWS, analysis of frontline banking interactions and engagement with financial institutions – to translate conceptual findings into systemic change. Without such engagement, communicative harm risks remaining invisible, normalised and unchallenged. Institutions must reframe the role of speech in evaluating credibility, legitimacy and compliance. By embedding neurodiversity into procedural design, interactional and communicative harms can be reduced and the inclusion of PWS can be more fully realised. When communicative diversity is recognised not as an exception but as a routine condition of social life, systems of trust and security become not weaker, but more just.