Judging Post-Controversy Expertise: Judicial Discretion and Scientific Marginalisation in the Courtroom

ABSTRACT

The sexual assault trial of R v Hartman included evidence from a sleep expert who found himself increasingly marginalised within the scientific community. Marginalisation takes place following a scientific controversy, when those considered to be on the losing side find it increasingly difficult to be heard by the community, and in particular, their ideas are removed from core texts in the field. Given a marginalised expert's ambiguous status, and a scientific knowledge deficit on the part of legal actors, on what grounds does a judge base their decision around the evidential value of their testimony? An analysis of the judge's decision in the trial indicates that she evaluated the expert's evidence by employing a version of a socio-technical review that included expectations of scientific rigour based on mechanical objectivity and procedural correctness. Drawing upon these processes and expectations of sound science, the judge had little difficulty evaluating the expert's evidence and finding it unsafe. In particular, she drew attention to the expert's mobilisation of a conspiratorial discursive style, a product of his marginalisation. This supports certain STS claims that legal actors already have tools for evaluating appropriate expertise, and these continue to be the cornerstone of judicial decision-making around expert testimony, even in highly ambiguous situations like post-controversy science.

KEYWORDS:

• Expert evidence
• judges
• marginalisation
• post-Controversy science
• sleep medicine
• alcohol-induced sleepwalking

Introduction

The presentation of expert evidence is understood as a routine part of criminal justice practice. While this has not always been the case (see for instance Crozier and Rees, 2012), it is now expected that criminal trials will generally include some form of expert testimony, most notably regarding forensic scientific or medical evidence. At the same time, the frequent mobilisation of scientific and medical expertise within legal proceedings has raised questions about judicial actors’ abilities to discriminate between legitimate expert claims and so-called junk science (i.e. technical claims that are not accepted by the scientific community). Junk science, the argument goes, is often presented by experts, willing to make weak scientific or fraudulent claims to win an adversarial trial out of their own self-interest (Hans and Saks, 2018). It is the role of the judge to decide whether an expert's testimony constitutes legitimate evidence that will be of value in resolving the trial. Scepticism over whether judges can make this distinction due to their scientific knowledge deficit (Hans and Saks, 2018), however, has resulted in many adversarial jurisdictions having introduced legal tests and scientific training to help judges distinguish legitimate from junk science (Caudill and LaRue, 2006; Jasanoff, 1995).

As part of observing and analysing the ways that scientific and technical expertise is constructed in legal contexts, Science and Technology Studies (STS) scholarship has been largely critical of the introduction of admissibility tests to enable judges to discriminate between divergent expert claims, variously asserting that: legal standards become rhetorical tools to allow a judge to present a decision, rather than being deterministic of said decision (Taipale, 2019a); the gatekeeping function turns judges into scientific knowledge producers (Jasanoff, 1995); and, the labelling of knowledge and practices as junk science is predominantly an exercise in boundary-work (Edmond and Mercer, 1998). Hence, consistent with STS work focusing upon non-legal domains, STS scholarship interested in legal trials has often emphasised the importance of the culture and practices shared by those expert witnesses and legal actors, sometimes referred to as the law-set (Edmond, 2001). Returning to the question of judicial decisions over the admissibility or validity of expert testimony, from an STS perspective therefore, it is these shared expectations and practices that serve as the basis upon which the credibility of scientific expert testimony comes to be judged.

The issue of the credibility of expert testimony leads to consideration of the nature of conflicting or controversial knowledge presented at trials. Adversarial systems are often characterised by experts for both the prosecution and the defence, who may not agree. When parties present conflicting claims, a judge must then discern between the two sets of expert interpretations of the evidence.¹ Divergent testimonies might be the result of accepted differences of opinion amongst experts working within a scientific community, or, less commonly, they might be the product of a broader scientific controversy that has extended into the legal sphere, wherein actors on both sides of that controversy try to secure alliances with legal professionals with the aim of achieving scientific closure (Golan, 1999; Lynch et al., 2008).

While judicial decision-making has been analysed with respect to both generally accepted (Edmond, 2001) and controversial science (Golan, 1999; Lynch et al., 2008), what has yet to be explored are the ways judges respond to expertise presented in the aftermath of a controversy, a time when consensus is forming around what ought to be the direction of knowledge, and those that disagree become increasingly marginalised (Wazeck, 2013). Following a trial that included expert evidence based on a controversial medical-scientific method and associated claims that had recently been undermined in the leading diagnostic manuals in the field, we examine how the judge in the case discerned the legitimacy (or not) of the expert's testimony. On what basis did she determine authoritative science, and how were the practices of the expert, who happened to be on the losing side of the controversy, understood in her decision-making responsibilities?

Concepts from STS are particularly fruitful for answering these questions as they emphasise the situated and contingent nature of expertise (Edmond, 2001; Taipale, 2019a). They are valuable as well for setting out and analysing processes of meta-expertise (Taipale, 2019b). In addition, STS scholarship that has explored post-controversy science, specifically the process of scientific marginalisation (Delborne, 2008; Wazeck, 2013), enables greater understanding of the ways in which marginalised narratives and behaviours might factor in judicial decision-making. We apply these analytical tools to a case study concerning expert testimony in a recent sexual assault trial centred on whether alcohol can trigger a sleepwalking behaviour. Drawing upon Wazeck’s (2013) finding that the apparent closure of scientific controversies can produce conspiratorial and victim discourses from scientists on the losing side, we demonstrate how discourses that run counter to the law-sets’ (Edmond, 2001) normative expectation of scientific practice may undermine legal actors’ trust in an expert and illustrate how a judge may draw on existing skills and knowledge to evaluate expertise.

Analytical Perspectives

Legal Meta-Expertise

The presence of scientific and medical experts in the courtroom can often result in a battle, with one legal team introducing their expert and the other side offering another to respond to or challenge the first's claims. Some scientific critics have raised concerns that judges do not have the necessary technical knowledge to adequately discriminate scientific from non-scientific forms of evidence, nor reliable from non-reliable forms of scientific or medical testimony (Hans and Saks, 2018). While those from a more constructivist philosophical or social scientific viewpoint do not necessarily endorse the viewpoint that judges have such a knowledge deficit, they have nonetheless recommended additional heuristics or training of judges in scientific forms of reasoning (Beecher-Monas, 2007; Haack, 2014; Jasanoff, 1995).

In contrast, Taipale (2019a) has suggested that simply improving judges’ understanding of scientific processes and reasoning is insufficient for increasing the reliability of judicial discrimination of scientific/medical evidence. Identifying a study assessing judges’ understanding and application of scientific standards included in legal tests, Taipale (2019a) notes that although judges received training in the scientific bases underpinning the legal tests, they were unable to fully understand and apply scientific standards. This finding was not surprising, as developing discriminatory scientific expertise is a craft practice, requiring apprenticeship within a scientific community (Taipale, 2019a). However, Taipale suggests that judges do in fact have significant expertise in legal processes and procedures (referred to above as the law-sets’ expectations and practices) and can discriminate between forms of scientific expertise based upon what he refers to as a socio-technical review of the expertise presented in the trial (Taipale, 2019b).

This judicial evaluation of expert evidence, according to Taipale (2019b), is based on both direct and indirect judgment of experts’ social and technical claims. In the absence of scientific expertise, he argues, judges rely on other proxies, including reviews of what the experts have said about each other, to help them form a conclusion. Key to the proxy measures that judges weigh in their review are the discourses that experts use to assist in the construction of their own credibility, or the deconstruction of the credibility of others. Credentials, experience, memberships in professional organisations and status within the relevant scientific community are all indicators that the expert is to be trusted. Alternatively, the appearance of an interest in the case, other than it just being a traditional part of the expert's work, or the sense that the expert is using the trial for an additional end can be interpreted as biases by legal actors, undermining the expert's credibility.

Taipale's concept of the socio-technical review draws heavily upon Edmond’s (2001) conception of the law-set. The law-set, which includes legal fact-finders and other legal actors (e.g. scientific experts), collectively co-constructs conceptions of legal-scientific acceptability. The socio-technical review includes many of the indicators Edmond highlights as being crucial for scientific acceptability for the law-set (credentials, experience, status with community etc.) with one important omission: scientific procedural correctness. Edmond (2001) notes that during trials, experts – and the legal teams introducing them – strategically attempt to present themselves and their testimony as credible by employing discourses that emphasise mechanical expertise or trained judgment (Daston and Galison, 2007; Edmond, 1999; Edmond, 2001; Lynch et al., 2008). During cross-examination, the opposing legal team can deconstruct those claims, a task that is often not overly difficult given the high standards of procedural correctness that experts claim to employ to assist their own credibility construction (Jones, 1994). If the expert is able to continue to convince legal fact-finders that their work was performed according to a high level (or at least an appropriate level) of procedural correctness, even after attempted deconstruction, then their credibility will have been maintained in the context of the trial.

Both Taipale and Edmond concentrate on trials where the legitimacy of the scientific knowledge being testified upon is not in question. While there may be controversy in the trial based on the differential interpretation of the evidence presented, there is no broader epistemological or scientific controversy amongst the relevant scientific community. The differences of opinion evidenced in the trial are the product of two (or more) authoritative experts legitimately interpreting the evidence differently. However, the case study we introduce below not only illustrates the engagement of a socio-technical review – inclusive of Edmond's procedural correctness – but raises an additional question: what happens in the aftermath of a scientific controversy? How does a socio-technical review take place during the post-controversy period when there is a significant consensus amongst many members of the community around how knowledge ought to develop (evidenced by these ideas being stabilised in authoritative texts, for instance professional manuals or textbooks), but some continue to disagree? To understand the context of the post-controversy period and its impact on the socio-technical review, we first must consider scientific marginalisation.

Marginalised Science and Scientists

Scientific marginalisation refers to the transition of members of a particular scientific community from legitimate claims-makers to being on the margins, or at worst, exiled or excluded. Marginalisation occurs following a scientific controversy when one group has been decided the winner of the controversy, their claims moving forward as scientific knowledge, while the losers become marginalised (see Collins 1985 for more on the closure of controversies). Marginalisation is not immediate but is a process and Delborne (2008) and Wazeck (2013) have set out similar procedures by which scientific actors become marginalised.

The first stage of the marginalisation process consists of a failure to accept the direction of the scientific community and continue to express claims or practices now deemed illegitimate/unauthoritative. This quickly translates to the second phase, wherein the same actors may continue to argue their case in the traditional scientific manner. Publications and letters to the editors of high-ranking journals sent during this phase will often be published alongside replies and rebuttals from other scientists highlighting supposed misunderstandings or errors on the part of the increasingly marginalised author(s). Alternatively, their correspondence might not be published at all and instead ignored, which Wazeck notes is the ‘most efficient form of marginalization’ (Wazeck, 2013, p. 167).

The community's lack of engagement with these scientists can lead to the third phase of marginalisation: the development of a defensive attitude and mobilisation of conspiracy discourses. Repeatedly having work, at worst rejected/ignored, or, at best published but with rebuttals can result in the marginalised scientist(s) taking a somewhat defensive and potentially paranoid stance. The scale of rejection they receive is considered wholly unjustified and only to be understood as the product of a conspiracy aimed at suppressing their work (Wazeck, 2013). They may increasingly mobilise discourses that present themselves as victims of a conspiracy to silence their work and imply a belief that once the conspiracy is over, and they are able to publish and be listened to again, they will be seen to be correct.

Given that they are not being published in the traditional scientific spaces, marginalised scientists may seek different arenas to express this conspiracy discourse, and so, in the fourth phase, the actors will often turn to other domains to generate alliances with organisations and institutions outside of the scientific milieu. This can result in associations with apparent political allies, for instance Non-Governmental Organisations (Delborne, 2008) or other political groups (Wazeck, 2013). The aim is to shift the debate away from the specific technical matters that produced the controversy in the first place to wider discourses about the nature of scientific work (e.g. freedom of speech in science publishing).

Delborne’s (2008) and Wazeck’s (2013) marginalisation processes offer detailed explanations for how scientific actors may transition from accepted members of a scientific community to marginalised outsiders. We extend these insights to include the courtroom as an arena for marginalised discourse, exploring a situation involving a scientific medical expert on the losing side of a particular controversy. Focusing on the judge's response to the expert's evidence, we examine the socio-technical review that she performed during a period of post-controversy science. To present this case study in a comprehensive manner, we first outline the alcohol-induced sleepwalking controversy, a contentious debate that arose in sleep medicine, the closure of which constitutes the background to the trial.

The Alcohol-Induced Sleepwalking Controversy

Sleepwalking is a disorder of arousal (DOA) that normally occurs during Non-Rapid Eye Movement (NREM) sleep and so is often identified as an NREM Parasomnia (American Academy of Sleep Medicine [hereafter AASM], 2014). DOAs involve incomplete awakening from deep sleep (the sleeper rapidly transitions from a deeper level of sleep to lighter) without a return to consciousness; technically labelled as arousals. Arousals can be accompanied by a range of physical movements, from walking around to driving vehicles (Schenck and Mahowald, 1995; Wolf-Meyer, 2012). Sleepwalking – as we will collectively refer to these activities, although some conditions have separate diagnoses, for instance Sleep Related Eating Disorder or Sexsomnia (Scheffer, 2010) – is considered to be caused by a combination of at least two of the following three factors: predisposing; priming; and precipitating (or triggering) factors (Pressman, 2007). Predisposing and priming factors relate to a combination of hereditary/genetic or health conditions and/or activities that serve as a background context in which a person's sleep can be disrupted. Crucial to whether a person will have a sleepwalking activity is the precipitating or triggering factor. This can include anything that might interrupt sleep and thereby result in an arousal, for instance sound or touch.

If the sleepwalking activity results in a criminal offence, the sleepwalker is entitled to an automatism defence, as, being asleep, they are not considered to be in control of their actions (Jackson et al., 2015). One of the key catalysts of the alcohol-induced sleepwalking controversy, was the production of a forensic test that attempted to identify whether alcohol had an effect on a person's sleep sufficient to trigger parasomnias, thereby enabling a potential sleepwalking/automatism defence (see also Ebrahim et al., 2005; Wilson et al., 2005; Jackson et al., 2015; Rumbold and Wasik, 2015).

The new forensic test was an adaptation of the traditional clinical sleep investigation, polysomnography (Kroker, 2007), by adding an alcohol provocation night into the usual routine of investigations (Ebrahim et al. 2005). Alcohol provocation, often known as alcohol challenge investigation, necessitated that the patient consume alcohol before falling asleep in order to record the impact of alcohol upon the person's sleep architecture. The person's sleep would then be video recorded and a range of additional measurements taken throughout the night. Ideally, evidence of an actual parasomnia would be the best evidence of someone having a history of parasomnias, but it is accepted that these happen rarely in clinical environments as they are temporally limited and often are highly dissimilar to a person's usual sleeping environment (Pressman, 2007). In the absence of full physical manifestation of parasomnias, the sleep experts Drs Ebrahim and Fenwick recorded the frequency of arousals from sleep, especially in the deeper periods of sleep (often referred to as Slow-Wave Sleep (SWS)), assuming a correlation between a high frequency of arousals from SWS with a likelihood to have parasomnias. If alcohol was shown to have an effect on a person's sleep pattern, for instance producing a higher rate of arousals from SWS, it could be important for informing the expert's report about the potential causes of the violent event (Ebrahim, 2006).

The alcohol challenge test was used in two English cases, R. v Lowe (Ebrahim et al., 2005) and R. v Bilton (Ebrahim, 2006) and the publication of the case reports for these cases within the forensic medical literature sparked a significant controversy in sleep medicine with one group of sleep experts, associated with Dr Mark Pressman, becoming deeply concerned about the development of the alcohol challenge test and the claims that alcohol could trigger a parasomnia. During the next five years a heated correspondence and publication record grew between Pressman (and colleagues) and Ebrahim and Fenwick, debating the legitimacy of alcohol challenge as a clinical or forensic test, whether arousals from SWS was an indicator of an underlying parasomnia and whether alcohol should be considered a trigger factor for parasomnias (see for instance Pressman et al., 2007a). Pressman and his colleagues accepted that the suspects did not recall the violent incidents of which they had been accused. However, they proposed that, rather than these incidents being the product of sleepwalking, the suspects had been so inebriated that they had entered an alcohol-based blackout and so had been unable to form memories of their actions (Cramer Bornemann, 2013). As both alcohol blackout and parasomnias involve activities later forgotten, alcohol blackouts can be misidentified as sleepwalking and misdiagnosed, as the person is not in fact asleep at the time (Pressman, 2013). Identifying a behaviour as a parasomnia event results in a legal defence, whereas if it is identified as solely due to intoxication, for which the accused was initially responsible, they are less likely to be able to form a defence around their lack of consciousness.²

Between 2007 and 2013 Ebrahim, Fenwick, Pressman and other sleep experts published various papers and letters justifying their own positions. While Pressman and colleagues often requested the retraction of Ebrahim and Fenwick's publications (see for instance Pressman et al., 2007b) on the basis that the alcohol challenge was junk-science and should not be used in legal contexts, they were unsuccessful. However, towards the end of this period, Pressman drew around himself and his group a large international network of sleep experts who were willing to criticise Ebrahim and Fenwick and their claim that alcohol constituted a trigger. For instance, Pressman et al.’s (2013) Letter to the Editor of Brain: A Journal of Neurology in response to a previous letter from Ebrahim and Fenwick (2012) included seven co-signatories, including sleep experts from Australia, Canada and the United Kingdom, as well as from the United States. Pressman et al.’s (2013) letter included a statement setting out the ‘generally accepted scientific position on the relationship between alcohol, sleep and sleepwalking’ (Pressman et al., 2013, p. 229) and argued that ‘there are no compelling scientific research data that support the notion that a reasonable amount of alcohol will either prime or trigger such an [sic] mixture of states as sleepwalking or sleep sex’ (Cramer Bornemann and Mahowald, cited in Pressman et al., 2013, p. 229). This statement, although contested by Ebrahim, Fenwick and others, would go on to form the basis for the updates to the parasomnia sections of the two chief diagnostic manuals in sleep medicine.

In 2013 and 2014, the Diagnostic and Statistical Manual Fifth Edition (DSM-V) and the International Classification of Sleep Disorders Third Edition (ICSD-3), were revised to state clearly that parasomnias should not be diagnosed in the presence of prior alcohol intoxication. The parasomnia section of the DSM-V would include a section on alcohol-induced blackouts in the differential diagnosis section that stated:

Alcohol-induced blackouts may be associated with extremely complex behaviors in the absence of other suggestions of intoxication. They do not involve the loss of consciousness but rather reflect an isolated disruption of memory for events during a drinking episode. By history, these behaviors may be indistinguishable from those seen in NREM sleep arousal disorders (American Psychiatric Association, 2013, p. 403).

The ICSD-3 was updated to state:

There is no scientific evidence that complex behaviors occurring during the sleep period following alcohol ingestion are anything other than the nocturnal wandering of an alcohol-intoxicated individual. Unconsciousness, intoxication, and sleep are very different states of consciousness. Parasomnias should be easily distinguished from alcohol intoxication by the presence of significant alcohol ingestion prior to bedtime (AASM, 2014, pp. 277–278).

The updates to the ICSD-3 and DSM-V were consistent with Pressman and colleagues’ position on the alcohol-induced sleepwalking controversy, and not that of Ebrahim and allies. The inclusion of Pressman and colleagues’ argument in the diagnostic manuals, which constitute the authoritative and legitimate knowledge of the sleep medicine community, putatively situate them as the de facto victors of the controversy.

The manual updates were not easily accepted, however, and shortly after publication another prominent sleep expert, Prof Rosalind Cartwright, was quick to criticise the updates, noting the personnel involved in their development:

The parasomnia working group [of the ICSD-3] had no member from another country [other than the United States of America], and three of the members frequently publish together. This homogeneity is, I believe, responsible for the one-sided, premature conclusion that excludes all alcohol-related arousals from being given a NREM parasomnia diagnosis in the absence of any hypothesis-testing studies supporting the decision (Cartwright, 2014, p. 1039).

Pressman and Cramer Bornemann (2015) responded by accepting that there was no definitive evidence but added that the overwhelming quantity of literature on the impact of alcohol on nocturnal wandering indicated that it was the conservative response to infer that alcohol alone could explain the violent act. They also responded to Cartwright's suggestion that the authors of the ICSD-3 update were biased and did not seek alternative opinions, outlining the various procedures and fora they provided for feedback and dialogue. They argued that the parasomnia section ‘was the result of the best international collaborations using best practices’ (Pressman and Cramer Bornemann, 2015, p. 188).

The sexual assault trial of R. v. Hartman in Ontario, Canada occurred two years after the publication of the ICSD-3, and the updates from that manual and the DSM-V played a major role. Dr Colin Shapiro, the sole sleep expert in the trial, was a collaborator of Ebrahim (see for instance Ebrahim et al., 2013) and, like Ebrahim, Fenwick and Cartwright, he remained convinced that alcohol constituted a trigger for parasomnia. As we will illustrate, Shapiro's response to the updates of the ICSD-3 and the DSM-V, in particular his mobilisation of a conspiracy discourse, had a significant impact upon the judge's socio-technical review in the case.³

Methods

This paper developed out of a broader project tracing the controversy around the alcohol-induced sleepwalking defence in forensic sleep medicine, particularly as it related to sexual offences and sexsomnia. Towards the end of that project, we became aware of a recent Ontario-based sexual assault trial and were interested in how the controversy would play out in the courtroom. We were curious as to the ways in which the judge would evaluate the expert's testimony given the close proximity to the tentative closure of the controversy. To ascertain how the judge evaluated the expert evidence in this trial, we systematically analysed her published judgement decision (Moore, 2018). While this is far from an ideal representation of the trial as it is a partial construction based on those aspects the judge intended to emphasise in her decision, as the space where the justifications for including or ignoring expert evidence are set out, makes this source wholly suitable to answer our questions (Edmond 1999, 2001; Taipale and Hautamäki, 2021). Like Taipale (2019b) therefore, we have analysed the written decision of a judge, in this case Judge Kimberly Moore, to assess how she interpreted the expert's evidence and incorporated it into her decision process. Unlike Taipale, our case study illustrates the performance of socio-technical review in the context of a single expert who appears to have been caught up in post-controversy scientific medicine.

In conducting this study, both authors independently read the judge's decision and developed a selection of major and minor themes in accordance with Framework Analysis (Ritchie and Lewis, 2004). We discussed and reflected upon these themes, refining the eventual list. All quotations used in the analysis are from Moore's decision and exemplify the evidence she wanted to emphasise in her judgement.

Before turning to the case study itself, it is important to touch upon the role of the controversy in this work. We approached our analysis symmetrically (Bloor, 1991; Delborne, 2008; Rees, 2019), not taking a position on the truth or falsity of the outcome of the scientific controversy. We are not sleep experts and, as such, do not have the contributory expertise (Collins and Evans, 2007) to assess whether alcohol does trigger a sleepwalking behaviour. Consensus statements have been published in the DSM-V and ICSD-3 to state that it does not. We have maintained an agnostic position as to whether this was the correct decision, and instead have focused on the possible impacts of these statements and being on the losing side of the controversy for this particular judicial socio-technical review.

Empirical Analysis

The Case of R. v. Hartman

In an Ontario court in April 2019, a sexual assault trial against Ryan Hartman ended in a rape conviction and a one-year sentence for the accused. The defence mobilised in the case was a parasomnia defence, and the trial addressed the question of ‘whether the appellant was NCR-MD [not criminally responsible due to mental disorder] at the time of the assault or guilty as charged’ [228].⁴

The uncontested facts of R. v. Hartman suggest that during a night of heavy drinking and socialising at the home of an acquaintance, Hartman climbed onto an air mattress on which the victim and her boyfriend were sleeping, and sexually assaulted her. Upon waking, the victim informed her boyfriend and others in the house of what had happened, and soon afterwards called the authorities. Hartman, upon arising, confessed to others, including the victim's boyfriend, that he woke up and found himself ‘making out’ [190] with the other man's girlfriend. It was also established that the accused had consumed 21 alcoholic drinks over the course of the evening.

In the spring of 2017, Hartman's representation mobilised a defence of automatism that was rooted in both a particular family history and evidence of prior sleep masturbation. Support for this defence came from: family members, who testified to experiences of sleepwalking behaviour in his youth, indicating a predisposing factor; his fiancée, who testified to having witnessed episodes of Hartman masturbating during sleep; and the accused himself, particularly regarding what others had told him of earlier sleepwalking behaviours.

Hartman also took part in a 48-hour sleep study conducted by Dr Colin Shapiro, at the request of the defence team. During the first night of sleep, he was observed and video recorded. It was noted that he demonstrated a 4–5 second episode of sexsomnia (masturbation), an arousal of which he was not aware and from which he fell back to sleep. The following day he took naps, and on the second evening, Hartman was instructed to consume alcohol prior to that night's sleep (an alcohol challenge investigation). An incident of bedwetting occurred on the night of the alcohol challenge, which Dr Shapiro interpreted as a parasomnia. In the judgement, Judge Moore noted that Shapiro testified that arousals from SWS could be triggered by ‘sounds, touch, sleep apnea, alcohol and drugs’ [183], which could then lead to parasomnia.

As a result, Hartman appeared to have good grounds for lodging a parasomnia defence: a family history, childhood sleepwalking, more recently observed sexual behaviour during sleep (including a recorded episode), clinically recorded high rates of arousals from deep sleep, and a noted parasomnia, specifically after the consumption of alcohol. Given this weight of evidence supporting his sleepwalking defence, why was Hartman found guilty?

Judge Moore's Decision

Judge Kimberly Moore positioned Dr Colin Shapiro's expert testimony as integral to her conclusion, especially as the grounds for the defence were largely tied to his expert report and Hartman's polysomnography (PSG). Judge Moore commenced in a highly favourable tone, setting out Shapiro's significant experience and identifying his expertise in ‘psychiatry, neuropsychiatry, and sleep disorders, including parasomnia and sexual behaviours that occur in the context of sleep disorders’ [179]. She also noted his long record of relevant publications, editorial positions, membership in various professional associations, directorships of sleep research related clinics and history of testifying before courts. On the face of it, Shapiro brought to the trial an abundance of attributes that enabled the construction of his credibility by legal actors (Edmond, 1999; Taipale, 2019). In the end, however, these characteristics proved insufficient for his evidence and testimony to remain credible in the context of this trial. We have identified three themes reflected in Judge Moore's decision to dismiss Shapiro's expertise: clumsiness and procedural irregularities; assumptions and relevance decisions; and conspiracy discourses.

Clumsiness and Procedural Irregularities

A key rhetorical tool used by experts when attempting to construct their own expertise and generate trust from legal triers-of-fact is to draw upon assertions of procedural correctness and mechanical or administrative objectivity (Daston and Gallison, 2007; Edmond, 1999; Edmond, 2001; Lynch et al., 2008). Judge Moore noted that Dr Shapiro was no exception to this practice. The decision indicated that in his testimony Shapiro went to great lengths to present himself as having been procedurally thorough, underscoring at several points his own adherence to mechanical objectivity and his trained judgment. For instance, he set out in detail the procedure for the PSG that Ryan Hartman underwent in his clinic, and when asked how he had determined a diagnosis, he emphasised the care and attention necessary in evaluating various forms of information. As stated in the judge's reasoning:

In examination-in-chief Dr. Shapiro explained that he carefully considered all of the available information, and did not come to his conclusion based on one piece of information. In cross-examination he advised the Court that as a clinician he does not simply accept what people say, if there is a reason to question the veracity of the information … he also confirmed that he looked at all of the evidence that supported his opinion, and considered whether there are other things that should be considered and excluded [203].

Indeed, further in the transcript, Judge Moore included Shapiro's own description of his work: ‘My approach is to try and be systematic and to get the clinical story’ [207], noting that he went on to state that it takes time and work to explore all evidence scrupulously.

Shapiro's initial claims of procedural rigour and care, however, were challenged through cross-examination. Judge Moore repeatedly highlighted information provided in his report and/or testimony that had been shown by the defence to be inaccurate, inconsistent, or contradictory to the evidence provided by the witnesses on whom he had reported. For example, rather than systematically collecting and engaging with the materials required for his expert assessment, the judge suggested that Shapiro was often unclear as to whether he had thoroughly read relevant reports, as he had stated at one point that he ‘skimmed over’ [212] notes about the case. He had also failed to correctly record the details of events such as the dates of the meetings he had had with Hartman, leading Moore to write: ‘Was Dr. Shapiro careless as to how he recorded that information? I do not know. Regardless of the reason for this significant discrepancy, I find it concerning’ [275]. Eventually, she concluded:

I found Dr. Shapiro's report very difficult to read, and very incomplete … I found Dr. Shapiro's evidence that he may have merely skimmed over these [notes from Hartman's previous lawyer] in his review of the file concerning. I found it even more concerning that, even when faced in court with the contents of these notes, he did not see that there were contradictions [290].

Contrary to Dr Shapiro's presentation of his work as rigorous, detailed and procedurally accurate, cross-examination and the combining of his testimony with those of the other witnesses revealed inaccuracies and errors that ultimately undermined, in the judge's estimation, the image of expertise that Shapiro had attempted to present.

Assumptions and Relevance Decisions

Judge Moore indicated that a pillar of Dr Shapiro's argument for Hartman's innocence related to the accused's actions following his supposed awakening from sleep following the sexual assault. Shapiro had interpreted Hartman's confession to the victim's partner as an act of trying to understand the information and reactions confronting him after waking up. This ‘filling in the blanks’ [209], according to Shapiro, was consistent with a person waking from sleep. He then went further, suggesting that struggling to comprehend would likely not be congruent with the actions of a rapist who, instead, would be quick to dismiss any accusations of impropriety:

There clearly is the possibility [Hartman is guilty] and the reason that he's in court is that it could’ve been some sexual assault. The point that I think makes that much less likely is that he was very quick to say to the person's, who accuses him, then boyfriend, who was in the bed at the same time, ‘I think I was making out with your girlfriend.’ People who commit rape don't volunteer themselves as having done something which has huge consequences. I think there was a lot of other factors at play, as well [190].

In her decision, Judge Moore emphasised that interpretations of this kind were beyond Dr Shapiro's expert qualifications in this legal context. She noted that his generalisations ‘made about how rape victims and sexual offenders would behave’ [279] caused her concern.

The judge also had misgivings about contradictions with respect to the factors Shapiro considered relevant to his decision-making. For instance, she highlighted that during cross-examination Shapiro was asked whether it would be the case that if Hartman had awoken during the night, he would have fallen straight back into a deep sleep, consistent with that necessary for parasomnia. Shapiro responded:

Again, if you were fully awake, it would take you a while to go back to sleep, and you don't go immediately back into deep sleep. You go into light sleep and then medium sleep, and then back into light sleep and maybe more medium sleep, and then you might get into a deep sleep [292].

Given that during the trial Ryan Hartman claimed that at one point during that night he had awoken with a pain in his neck from falling asleep on an armchair and had moved from that chair onto the air mattress upon which the victim and her partner were sleeping, the defence engaged in the following exchange with Dr Shapiro:

Q. Would you have paid greater attention to it if when he had the kink in the neck, he got up and walked around? Would that have been more important as … ?

A. Then I would say he was awake, and it would be … it certainly would be relevant. And the likelihood of him going immediately back into deep sleep is less likely [197].

Subsequently, when the prosecution continued to press Shapiro on this matter, especially if the awakening and movement from the armchair to the mattress potentially negated a sleepwalking activity, as Hartman was not at a deep enough level of sleep, they had the following exchange:

Q. When he woke up with a kink in his neck, why would you ignore that piece of evidence?

A. It didn't seem salient to me.

Q. So because it didn't serve your purpose?

A. No.

Q. Why does it not seem salient?

A. If he’d said he dreamt about Batman in the middle of the night before, I might record that he dreamt about Batman, but I wouldn't not think it's very important in the general scheme of things. The fact that he had arousals during his sleep, that is relevant [200].

Dr Shapiro was comparing two forms of evidence here: the clinical evidence demonstrating that Hartman had high rates of arousals from sleep, especially following alcohol consumption; and the expert inference that sleep patterns are cyclical and a person is unlikely to move into a deeper level of sleep, the kind necessary for most parasomnia events, shortly after initially falling asleep. The clinical evidence was consistent with Shapiro's account of Hartman performing a sexsomnia activity, while the medical inference contradicted it. Referring to this contradiction in her decision, Judge Moore suggested that Shapiro demonstrated ‘an unwillingness to accept any evidence that contradicted his opinion that Mr. Hartman was suffering from a parasomnia at the time of the incident’ [294]. Clear hints as to the basis for such unwillingness can be found in the third theme emerging from the judge's decision – the conspiracy discourse.

Conspiracy Discourse

One of the most salient points to arise in the judge's decision concerned the current state of scientific consensus regarding the role of alcohol in relation to parasomnia. Into this line of interrogation was introduced the topic of Dr Mark Pressman, who, while not physically present at the trial, had prepared a report for an earlier appeal case concerning Hartman, which resulted in a retrial, the current case under analysis. In that report, Pressman concluded that Hartman was not asleep but intoxicated during the assault, consistent with his position that intoxication does not trigger sleepwalking (Doherty, 2015). Judge Moore noted that the prosecution asked Dr Shapiro if he had read Pressman's report, to which he responded that he had. The prosecution then proceeded to ask questions related to the relationship between intoxication and sleepwalking:

Q. The alcohol issue is, in your expertise, you believe that it's a trigger to parasomnia. Is that correct?

A. There's no question that it is a trigger to parasomnia.

Q. The difficulty is the DSM-5 says no.

A. Do you know who wrote the DSM-5 section on that?

Q. Well, is it not the bible that people follow?

A. If you go to the DSM-4, which was the previous bible, so now there's a new religion. [215]

Following this, Shapiro launched an attack on Pressman based on the role he had played in developing that portion of the DSM-V:

Dr. Pressman was very instrumental. He was the chair of that section. He has a particular view, which is a minority view … and he's a very forceful individual. So he has promulgated that view, and it now appears in DSM-5. My own view and the view of a lot of Canadian experts, is that he has hoodwinked the DSM committee [215].

At another point, as the prosecutor was introducing the ICSD-3, the following interaction took place:

Q: The parasomnia core group was chaired by Michel Cramer Bornemann and included members; Carlos Schenck, Mark Pressman, and Gerald Rosen. The process was quite comprehensive as follows.

A. Can I stop you there for a second? So I just want to point out that in terms of a pattern, Bornemann, Schenck, and Pressman have co-authored papers. I don't know about Gerald Rosen. So if those … if they’re four people on that committee, you know, what is striking is that there's a number of people who have an opinion that they know is at variance with the opinion that they held, and it's an important issue. None of those people, like Ebrahim from the UK or Moscovitch from Alberta, or myself or Gojer, were on those committees. That's fine. I mean, I don't need to be on a committee. But I’m just saying that it is a rigged jury, if you like, before it starts. [216].

Dr Shapiro's argument here repeated the claim earlier made by Cartwright that there was a core group who worked together and shaped the diagnostic manuals to their own way of thinking, ignoring the differing opinions of others. However, Shapiro's language was somewhat starker, situating those who worked on the diagnostic manuals as being ‘hoodwinked’ by Pressman, while those who did not agree with him were side-lined and ignored.

Judge Moore observed that Shapiro continued this conspiracy discourse the next time he appeared as an expert witness several months later in October 2017. This time the discourse was in response to the prosecution's referral to a statement from the DSM-V claiming that the number of arousals from SWS was not a strong indicator of sleepwalking as it was not a specific enough test. Dr Shapiro, at that time, had emphasised the importance of the presence of arousals as an indicator of a parasomnia and argued that the belief that they were not sensitive or specific enough was Dr Pressman's opinion, which he had imposed on the DSM committee:

No, no, he was part of the group, and he was the steering voice on that, and he's imposed his personal beliefs on to that, with two influential colleagues who’ve both come back from that position somewhat, that it's now enshrined. And until DSM-6 comes along, we will have that view as one view that is held. That doesn't mean that it's gospel [217].

Unfortunately for Shapiro, Dr Pressman was not a member of the committee for the Parasomnia section of the DSM, and at a March 2018 court appearance, after being confronted by the prosecutor about inaccuracies in his statements, ‘Dr. Shapiro acknowledged that he was incorrect about Dr. Pressman's role in the DSM-V’ [218]. Judge Moore recorded that ‘not only was Dr. Pressman not listed as the chair of that group, he was not even a member of the group’ [218].

Shapiro's claims about Pressman's manipulation of the updates to the diagnostic manuals did not go over well with Judge Moore or other members of the court. In fact, the judge noted that the defence team made a formal request that his comments regarding Pressman be disregarded. This request was denied, and this issue was reflected in the judge's decision:

It is certainly to be expected that expert witnesses will take contrary positions to other experts, and it is also expected that expert witnesses will be asked to comment on the findings of other experts. This is quite usual, and acceptable, as the whole point is to assist the Court in understanding issues that are outside the realm of knowledge of the Court … However, in this case I find that Dr. Shapiro's comments directed towards Dr. Pressman were unsolicited, gratuitous and non-responsive to the questions asked of him by [the Crown]. Most importantly, his criticism of Dr. Pressman's role in the DSM-V were completely unfounded and factually incorrect. I find that Dr. Shapiro was careless as to many of his comments about Dr. Pressman before this Court and that his comments demonstrated a lack of professionalism [281].

The Verdict

While the defence team mobilised a range of evidence in the trial, including testimony from family, bed partners and Hartman's own experience of sleep behaviours, they relied predominantly on Dr Shapiro as the sleep expert. The technical evidence presented to the court, including a visual recording of Hartman's sexual sleep behaviour as an indication of a parasomnia following the alcohol challenge PSG, was central to Shapiro's contribution.

Judge Moore, however, chose to disregard this evidence. She noted: ‘it would be unsafe for me to rely upon the evidence of Dr. Shapiro in this case. I find that Dr. Shapiro's evidence demonstrated bias and a failure, and an unwillingness, to consider all relevant information’ [278]. In identifying issues such as: Shapiro's ‘failure to adequately address the issue of alcohol in his report and the potential for alcoholic blackouts to have played a role in Mr. Hartman's memory loss’ [279]; his inadequate, unreliable and contradictory reading of the evidence with which he had been presented for his assessment; and what she concluded was Shapiro's ‘unprepared[ness] to accept any information that would change his view’ [290] in the face of his concluding narrative of what occurred, the judge suggested that he held biases towards a particular outcome. In contrast to what is expected of an expert, she stated: ‘I find that, at times, Dr. Shapiro was confusing his role as a witness, with that of an advocate’ [280]. With the expert evidence supporting Hartman's claims to having experienced an automatism undermined, Moore concluded that Ryan Hartman had experienced an alcohol blackout, not a parasomnia, and found him guilty.

Moore's judgement that Shapiro was acting as an advocate makes sense when understood in the context of the alcohol-induced sleepwalking controversy, the marginalisation process, and an extended version of Taipale's socio-technical review that includes procedural correctness. By introducing the controversy into the trial, the prosecution provided Shapiro with the space to explain and defend his position on alcohol-induced sleepwalking. In contrast to phase four of Wazeck’s (2013) marginalisation process, where the marginalised seek out others who will provide a friendly/sympathetic ear, Shapiro had no need to find others, as his credibility and reputation were still sufficient to provide him with access to the courtroom as an expert. Nevertheless, once the opportunity arose to discuss the controversy and the updates to the manuals, Shapiro was quick to both justify his own position (that alcohol triggered sleepwalking, and that this was the basis upon which he was assisting the defence) and engage a conspiracy discourse explaining that he and like-minded sleep experts had been unable to receive a fair hearing in sleep medicine because Pressman had manipulated the community.

Situating Pressman as the master manipulator and chief villain, Shapiro appeared to be very much mobilising Wazeck’s (2013) conspiratorial discourse, emphasising the victimhood of both himself and his colleagues. Following Delborne (2008) and Wazeck (2013), we suggest that such discourses are the frequent result of the post-controversy process and thus perhaps Shapiro could not help but marshal victimhood and conspiracy narratives when provided with the opportunity. Unfortunately for Shapiro, the ways that legal and scientific actors, the law-set (Edmond, 2001), have collaboratively developed normative expectations around the performance of expert evidence, the sense that a trial is being used by an expert witness to further their own interests or ends is seen as bias. Shapiro's case, however, identifies that this bias was not as easily reduced to that of a simple economic interest, but rather a well-respected and well-credentialed medical scientist who happened to support a particular standpoint that was losing legitimacy within the sleep medicine community.

Conclusion

This article focuses on a recent sexual assault trial centred on a defence of alcohol-induced sleepwalking as a means by which to examine the role of expert medical scientific evidence and testimony in the judicial decision-making process. This case involving a single expert who was increasingly marginalised within the sleep medicine community provided the opportunity to ask: how did a judge evaluate the contributions of an expert in the aftermath of a scientific controversy related to the case? Drawing upon Taipale’s (2019b) socio-technical review and Edmond’s (2001) concept of procedural correctness, as well as Delborne’s (2008) and Wazeck’s (2013) conceptualisation of marginalised scientists, we examined the judge's process and reasoning in assessing the expertise presented in determining a verdict.

Taipale's socio-technical review indicates that a judge may use a variety of proxies including direct and indirect judgments of the technical and social claims (including credentials and standing in the field) of experts, as well as what one expert may say about another to bolster their own credibility or undermine that of the other. Our analysis, based on three themes reflected in Judge Moore's decision, illustrates the particular ways in which her socio-technical review reflected both the Taipale construct, as well as Edmond's useful identification of the importance of procedural correctness. Alongside this, our case study also noted the difficulties that marginalised experts might have in the courtroom. Despite acknowledgement of Shapiro's credentials and professional status, the procedural irregularities, inappropriate and irrelevant assumptions, mobilisation of a conspiratorial discursive style and the appearance of having a personal interest in the case led Judge Moore to clearly articulate that she had no difficulty removing Colin Shapiro's testimony from her judicial reasoning, undermining the potential probative value of his evidence.

Although the basis for Moore's justification for the disregard of Shapiro's testimony was his bias and advocacy, close scrutiny of the case identifies that his interest was not simply economic; on the contrary, he was well-credentialed and well-respected in the community but was testifying in support of a position that was losing favour within sleep medicine. His marginalisation, evidenced by his opposition to the state-of-the-art found in medical manuals, resulted in him responding to cross-examination in ways that did not conform with the normative expectations of the law-set. Judge Moore was therefore able to identify Shapiro's clear interest in the case and found him not performing the role of the independent expert.

This study contributes to the STS literature by being the first to link marginalisation/post-controversy science and legal meta-expertise. Even the most authoritative expert can be reduced to mobilising conspiracy discourses if they happen to find themselves on the losing side following the end of a controversy. If such a person ends up providing expertise in a trial, their articulation of conspiracy discourses will not conform to the idealised images of expertise that the law-set have collectively developed. Experts like Shapiro are not selling their expertise with little integrity but rather find themselves in a complex environment, defending positions that are becoming increasingly untenable within the scientific community but relying on discourses that make them increasingly unlikely to be believed.

While Edmond's law-set and Taipale's socio-technical review have been used successfully in the non-controversy science context, we believe that they are also particularly salient when coupled with post-controversy science and marginalisation as they provide a more sophisticated explanation for why experts might present as having an interest in legal contexts; i.e. they highlight the interactional dynamics between the difficulties experts have challenging a forming consensus, once it is formalised within key texts, and the ways that the courts and other bodies defer to these core texts. Moreover, while we do not have the contributory expertise (Collins and Evans 2007) to evaluate whether Moore was correct to disregard Shapiro's evidence, this case study suggests that contemporary anxieties around judicial knowledge deficiencies and attempts to improve judges’ discrimination of expertise should look to the existing skill-sets that judges have collectively co-constructed with the scientific community, and build from them, rather than imposing scientific practices that judges have little understanding and experience of. Evaluating testimony in this way enables judges to concentrate on their own expertise and thereby provides the opportunity to focus on what is really at stake, the real-life outcomes for criminal trials and those involved.

Acknowledgments

The authors wish to thank Prof Levidow, Prof Birch and the two anonymous reviewers, as well as Prof Barbara Prainsack and Dr Erik Borve Rasmussen, for reading earlier drafts and providing insightful and supportive comments. As always, any errors, omissions or inaccuracies are solely our own.

Disclosure Statement

No potential conflict of interest were reported by the author.

Additional information

Notes on contributors

Gethin Rees

Gethin Rees is a Senior Lecturer in Sociology at Newcastle University. He is interested in the intersection of healthcare and criminal justice, whether embodied in healthcare professionals working in criminal justice contexts (e.g. police stations and prisons) or in scientific and/or medical experts presenting evidence in criminal trials.

Deborah White

Deborah White is a Professor in the Department of Sociology at Trent University in Ontario, Canada. Her research focuses on the institutional responses to sexual violence, particularly medico-legal interventions and the role and nature of forensic evidence and experts in criminal justice systems. As a feminist scholar of Science and Technology Studies, she also conducts critical research on technologies of rape and sexual assault, specifically anti-rape technologies.

Notes

1 This is most predominant in adversarial jurisdictions where the defence and prosecution introduce competing experts, although increasingly experts are being asked to produce joint expert statements for the court rather than appearing for a particular side in the trial, more akin to inquisitorial trials (Ebrahim and Fenwick, 2007). However, as Bal (2005) and Taipale (2019b) have identified, even in inquisitorial jurisdictions, competing experts are often employed.

2 It is not always the case that intoxication cannot be used as a legal defence; see for instance Simester (2009). Additionally, the Ontario Court of Appeal have recently reiterated that voluntary intoxication to the point of automatism can constitute a defence (DiManno, 2020).

3 While the updates to the ICSD-V and the DSM-3 constitute an enactment of closure of the alcohol-induced sleepwalking defence controversy, this does not necessarily mean that the debate will not open up again. Consistent with the claims of Finitism (Barnes et al., 1996; Bloor, 1991; Rees, 2011), new cases, findings or interpretations can always problematise a consensus, and so all claims are open-ended and revisable. At the time of writing, a new article in the Journal of Forensic Sciences takes to task the authors of the manuals for flawed statistical analysis (Munro, 2020). It is presently too early to know what impact Munro's intervention will have on the sleepwalking defence and legal expertise, but the case we have used in our analysis occurred shortly after the publication of the manuals, and as we go on to show, they indicated, at least as far as the court was concerned, the state of the art of medical knowledge.

4 All citations refer to Judge Moore's decision (Moore, 2018) and so we identify each quotation by the paragraph number from which it was taken.