The Moorov doctrine and coercive control: Proving a ‘course of behaviour’ under s. 1 of the Domestic Abuse (Scotland) Act 2018

In 2019, a distinct offence of ‘abusive behaviour towards partner or ex-partner’ (‘domestic abuse’) came into force in Scotland via s. 1 of the Domestic Abuse (Scotland) Act 2018. This new offence has been celebrated for its meaningful incorporation of the concept of coercive control (Evan Stark has described the 2018 Act as ‘gold standard’ legislation) and may serve as a model for other jurisdictions looking to criminalise coercive and controlling behaviours. The practical effectiveness of the offence in Scotland, however, will hinge on how Scotland’s corroboration rule, and the accompanying Moorov doctrine (‘Moorov’), are applied in this context. Drawing both on recent doctrinal developments and on a conceptual understanding of the dynamics of coercive control, this article offers the first in-depth analysis of how Moorov is likely to apply in s. 1 cases. It identifies developments that are likely to assist the prosecution, as well as potential barriers to the doctrine’s successful application, and argues that in certain cases judges and jurors will have difficulty seeing the ‘course of conduct’ required by Moorov without proper understanding of the policy underpinning the Act and the gendered nature of domestic abuse. The article considers how this understanding may be brought about, both within the confines of the current law and in terms of possible reform.


Introduction
It is increasingly recognised that overcoming evidential challenges will be key to ensuring the effectiveness of the recent global wave of legislation criminalising coercive and controlling behaviours (Bettinson and Bishop, 2018;Burman and Brooks-Hay, 2018). In Scotland, where a distinct offence of domestic abuse modelled on the concept coercive control was recently introduced via the Domestic Abuse (Scotland) Act 2018, such concerns are especially pronounced due to the existence of a formal legal requirement of corroboration. 1 This rule requires that the 'essential' or 'crucial' facts of a case-including that a crime was committed and that it was the accused who committed the crime-are supported by two independent pieces of evidence, and therefore raises obvious issues in cases that tend to occur in private (Cairns, 2013). Although controversy over the way in which this requirement operates in domestic abuse cases predates the 2018 Act, securing corroborative evidence for the purposes of the new offence presents new complexities both due to the fact that there are likely to be fewer sources of evidence (e.g. forensic evidence, CCTV or direct witnesses) available to corroborate the newly criminalised coercive and controlling behaviours, 2 and due to the precise way in which the terms of the legislation interact with the technical aspects of the corroboration rule.
The Scottish doctrine of mutual corroboration-the Moorov doctrine 3 -alleviates the inherent evidential challenges in this area to a certain extent, and its successful application will be pivotal to securing prosecutions and convictions under the 2018 Act. The doctrine is not an exception to Scotland's corroboration requirement, but rather one of several rules that has developed to assist with securing corroborative evidence in circumstances where it is likely to be lacking, including in sexual offence and domestic abuse cases. 4 This article offers an in-depth doctrinal analysis of recent developments in the interpretation of the Moorov doctrine and their relevance for proving the Scottish offence of domestic abuse, which came into effect in April 2019. In addition to providing an up-to-date account of the precise factors and considerations that are relevant to establishing a course of conduct under Moorov, and an assessment of how these are likely to apply in the context of proving the distinct offence of domestic abuse, this article also draws on case law to identify several potential stumbling blocks to the doctrine's successful application in s. 1 cases. These include, but are not limited to, the requirement for compelling similarity between offences where there is a lengthy time gap (explored in Part two); resistance to further extension of the doctrine (explored in Part three); and the challenges of ensuring that both judges and jurors have sufficient understanding of coercive control-a concept that strongly influenced the creation and drafting of the 2018 Act-to be able to see connections between offending behaviour that has traditionally been viewed as different in nature (discussed in Part four). This article argues that judicial and juror understanding of the dynamics and gendered nature of coercive control is essential if the Moorov doctrine is to meet its full potential in this context; identifies ways the current law both facilitates and impedes this understanding; and considers possible avenues of reform. While there are a number of complex evidential issues raised by the new Scottish offence-including, for example, the use of hearsay evidence 5 and proving intent or recklessness to cause harm without the complainer's presence in court-this article concentrates exclusively on the practical application of the Moorov doctrine. It does not seek not provide a comprehensive analysis of all aspects of the 2018 Act (for this 1. In 2011, the Carloway Review (Scottish Government, 2011 recommended that the corroboration requirement should be abolished, triggering legislative reform plans and a heated debate. These plans were put on hold in 2015, following the publication of the Post-corroboration Safeguards Review (Scottish Government, 2015). For analysis of this U-turn see Cairns (2018). 2. This was acknowledged by the Crown Office and Procurator Fiscal Service (COPFS) as the Domestic Abuse (Scotland) Bill passed through the Scottish Parliament, but was accompanied by the reassurance that investigators and prosecutor s have the requisite 'skills and expertise' in this area to ensure that prosecutions still go ahead (Scottish Parliament, 2017). See also Burman and Brooks-Hay (2018: 8-9). 3. Moorov v HM Advocate 1930 JC 68;1930 SLT 596. The doctrine of mutual corroboration predates Moorov (it is referred to in the work of both Hume and Alison). However, Moorov was a landmark case that set out in detail the principles of mutual corroboration. At points in this article, the Moorov doctrine will simply be referred to as 'Moorov '. 4. For analysis of the other rules see Duff (2012). 5. For recent discussion of the use of hearsay in domestic abuse cases see Bettinson and Bishop (2018: 17-21). see Burman and Brooks-Hay, 2018), nor to address the difference that abolition of the corroboration rule would make in terms of improving prosecution and conviction rates in so-called private-sphere crimes (this has been explored in past work: see Cairns, 2013 and. Although focused on Scots law, this article contributes to an emerging global conversation about the complexities of evidencing coercive and controlling behaviours and engages with a range of issues that are of more general and international interest, including: the relationship between corroboration requirements and domestic abuse/sexual offending, the fluidity and indeterminacy of the concept of similarity, and the respective roles of judge and jury in assessing the sufficiency of evidence. It also contributes to the vast literature on the inherent limitations of both legislative and criminal justice responses to genderbased violence (for a recent example see, e.g., Cowan, 2019). Moreover, through offering the first comprehensive analysis of the specific evidential difficulties involved with proving the new Scottish offence, the article illustrates that the precise evidential challenges of proving distinct offences of domestic abuse or coercive control will vary from jurisdiction to jurisdiction. In the process of analysing the relationship between s. 1 of the 2018 Act and Moorov, this article makes a number of more general observations-which are particularly relevant to the prosecution of sexual offences-about the extension of Moorov by the High Court of Justiciary (hereinafter 'High Court'). Taken together with the analysis of Moorov and s. 1, these observations raise significant and timely questions about the robustness of the Moorov doctrine and, by extension, the Scottish corroboration requirement as a whole.
Part one: The natural connection between the Moorov doctrine and s. 1 of the 2018 Act The Moorov doctrine is applied on a day-to-day basis in the Scottish courts and is the closest that Scots law has to a doctrine of similar fact evidence. 6 As Lord Justice General Clyde set out in the case of Moorov itself, the doctrine allows mutual corroboration between separate acts where they are libelled together, there is a connection in 'time, character and circumstance', and they are bound together by a 'unity of intent, project, campaign or adventure.' 7 A more modern account of the law was recently provided by the Lord Justice General Lord Carloway in the 2019 case of HM Advocate v SM(2): It is not enough simply to catalogue some similarities between two crimes, and to dismiss others, for mutual corroboration to apply. There requires to be an overall similarity in the conduct described in the offences such as identifies it not just as constituting separate criminal episodes, but as 'component parts of one course of conduct persistently pursued by the accused.' 8 The requirement for a course of conduct under Moorov ensures that 'isolated instances of similar conduct' 9 are ineligible for mutual corroboration and therefore acts as a gatekeeper against the admission of propensity evidence that may cause prejudice to the accused. In early 2020, the High Court went as far as to describe the requirement for a course of conduct as a 'procedural safeguard'. 10 That said, some would no doubt take the view that the liberal interpretation of the Moorov doctrine is blurring the line between course of conduct evidence and general propensity evidence to an unacceptable extent. For example, in his recent analysis of similar fact evidence and Moorov, Davidson (2018: 102) observed that 'It may be thought . . . that Scots law is tending more and more towards the view that the fact that two or more individuals separately accuse an individual of the same broad category of crime is inherently 6. For analysis of the relationship between similar fact evidence and Moorov see generally Davidson (2018); Duff (2002 probative.' Controversy surrounding the extension of the Moorov doctrine is explored in more detail in Part three of this article. The usefulness of the Moorov doctrine in helping to prove domestic abuse in the Scottish courts stretches much further back than the creation of a distinct offence. On an initial glance, however, Moorov appears to be particularly well-suited to the new offence because s. 1 requires the prosecution to prove a course of behaviour, defined in s. 10(4) as involving behaviour on at least two occasions. For this course of behaviour to constitute the offence of 'abusive behaviour towards partner or ex-partner' there are two further conditions that must be met: first, that a reasonable person would consider A's course of behaviour to be likely to cause B to suffer physical or psychological harm (s. 1(2)(a)) and, second, that A either intends that, or is reckless as to whether, the course of behaviour will cause B physical or psychological harm (s. 1(2)(b)). The definition of 'abusive behaviour' is contained in s. 2 and is deliberately wide in order to capture the range of harms that can be experienced in an abusive relationship. It includes behaviour that is 'violent, threatening or intimidating' (s. 2(2)(a)) and behaviour directed at B, B's child or another person that either has as its purpose/s the bringing about of 'relevant effects' or that a reasonable person would consider likely to bring about relevant effects (s. 2 (2)(b)). These relevant effects are: a. making B dependent on, or subordinate to, A; b. isolating B from friends, relatives or other sources of support; c. controlling, regulating or monitoring B's day-to-day activities; d. depriving B of, or restricting B's, freedom of action; e. frightening, humiliating, degrading or punishing B.
Significantly, violent behaviour includes both physical and sexual violence (s. 2(4)(a)). One notable feature of the new offence, and one that distinguishes it from other comparable offences, is that it captures both behaviour that was already criminal under Scots law (e.g. assault, threatening and abusive behaviour, sexual offences) and behaviour that was not previously criminal (i.e. psychological and/or emotional abuse).
Although the term 'coercive control' does not appear anywhere in the 2018 Act itself, the offence 'draws centrally' (Burman and Brooks-Hay, 2018: 68) upon Stark's work on coercive control which emphasises the 'interweaving' of 'repeated physical abuse with three equally important tactics: intimidation, isolation and control' (Stark, 2007: 5). The principal argument for the creation of a distinct offence was that existing criminal offences did not, in theory or in practice, adequately capture this type of ongoing abuse that has cumulative impact and is rooted in gender inequality. Fixing this 'disjuncture' (Douglas, 2015: 447) between social science research (most notably that of Stark) and the criminal law's 'narrow temporal lens and . . . limited conception of harm' (Tuerkheimer, 2004: 972) is the objective of criminalisation efforts in Scotland and elsewhere. 11 The problem with the narrow, individualistic focus of the criminal law is that it ignores the context in which single incidents occur and therefore obscures the dynamics of control, power and gender inequality that make coercive control distinctively wrong and harmful (Tuerkheimer, 2004;Tadros, 2004Tadros, -2005. Appreciating context is important because coercive control research illustrates that, while particular incidents may appear relatively innocuous to outsiders, or when viewed in isolation, they can have a very different and sinister meaning within the context of a coercive and controlling relationship. These actions or behaviours may be implicit but 'credible threats' (Dutton and Goodman, 2005), such as a particular look or the passing of a sweatshirt in front of others to convey the need to cover physical injuries later (Stark, 2007: 229). As discussed later in this article, identifying such actions or behaviours as abusive is made more challenging due to gendered and heteronormative assumptions surrounding what is normal or expected in a romantic relationship (the 11. Addressing in detail the arguments for and against criminalisation of domestic abuse, either or in Scotland or generally, is beyond the scope of this article. For this see Tadros (2004Tadros ( -2005 and Bettinson (2016).
gesture of a man passing a sweatshirt to his female partner, for example, may be misread by others as an act of kindness rather than as a threat) (Stark, 2007: 229). As recently explained by Bettinson and Bishop, the strategies of coercive control 'serve to reinforce a specific construction of feminine identity and, due to the cultural association of masculine identity with control, the male dominance that is typically seen in a relationship characterised by coercive control may be hard to discern' (Bettinson and Bishop, 2018: 9). With this background understanding of coercive control, the Scottish Government set out to draft legislation which would better reflect the lived experience of those in abusive relationships or, to use Moorov terminology, would allow a court to see the accused's 'unity of intent' and 'campaign' of abuse in its full context. The 2018 Act therefore gives the Crown the flexibility to libel the entire ambit of the accused's behaviour within a single offence (Scottish Government, 2018: 1). In common with the Moorov doctrine, the offence demands a focus on the underlying connection between seemingly individual incidents of behaviour. The criminalisation of a full range of abusive behaviour in a single offence under s. 1 can be contrasted with the offence of 'coercive and controlling behaviour' in England and Wales (s. 76 Serious Crime Act 2015), which, it is worth noting in passing, criminalised coercive and controlling behaviours that were not previously captured by the criminal law. 12 Despite its deliberate avoidance of the term 'coercive control' due to definitional complexities (Burman and Brooks-Hay, 2018: 73), the 2018 Act therefore more accurately reflects and incorporates the concept that inspired its creation, and has been described by the world-renowned expert on coercive control, Professor Evan Stark, as 'gold standard' legislation (Scott, 2020: 177) that represents 'one of the most radical attempts yet to align the criminal justice response with a contemporary feminist conceptual understanding of domestic abuse as a form of coercive control' (Stark and Hester, 2019: 85). Although the manner in which the legislation is drafted may be in line with Stark's gendered analysis of coercive control, the placing of the Scottish offence on a pedestal overlooks the evidential challenges specific to Scotland and the way in which s. 1 will interact with unique features and rules of our justice system. 13 Chief among these challenges is the practical application of the Moorov doctrine. The value of this doctrine in assisting with securing prosecutions and convictions under s. 1 cannot be overstated-without the doctrine it would be impossible to satisfy the corroboration requirement in a great number of cases.
The usefulness of Moorov in the domestic abuse context was strengthened significantly by the decision in HM Advocate v Taylor, 14 in which the High Court confirmed both that 'the principle of mutual corroboration does not depend to any extent upon whether the separate incidents on which reliance is placed are charged in separate charges or in one single composite charge' 15 and that Moorov can apply to charges involving the same complainer where one of the charges is corroborated. It is worth briefly detailing the facts of this case to illustrate how Moorov works in practice. Taylor had been found guilty at Aberdeen Sheriff Court of repeatedly exposing himself at the windows of his home and masturbating in front of the complainer, Ms SLD. The only independent piece of evidence was an account from the complainer's partner, Mr R, who had witnessed the fourth and final incident of the accused's behaviour. On appeal to the Sheriff Appeal Court, the accused's conviction was restricted to this final incident as, in the court's view, the other incidents were uncorroborated. This decision was 12. There are a number of other differences between s. 1 and s. 76 but discussion of these is outwith the scope of this article. For indepth analysis see Bettinson (2020). 13. Of course, the content of the legislation could be 'gold standard' in theory and in practice in jurisdictions with different laws of criminal evidence. 14. HM Advocate v Taylor  reversed by the High Court, which refused to accept the respondent's arguments that the offences had to have been committed against more than one person, and supported by evidence from more than one complainer, in order for Moorov to apply. Delivering the opinion of the court, Lord Glennie noted that there would have been no question over the applicability of mutual corroboration had Mr R been named as a complainer to the fourth incident, and that 'it can make no difference that he was not a complainer but simply a witness.' 16 Following Taylor, then, it is clear that Moorov can be used in order to establish an alleged 'course of behaviour' against a single complainer under s. 1, providing that at least one of the incidents is supported by relevant evidence from one more source. As many s. 1 cases will only involve a single complainer (the accused's current or most recent partner), this is a positive development from a prosecutorial perspective. Taylor also clarified that the setting out of the alleged incidents comprising a course of behaviour in a composite charge, as would likely be the case under s. 1, is not a barrier to the application of Moorov. Indeed, setting the incidents out in a composite charge may well assist with demonstrating a 'course of behaviour'. What remains less certain is how exactly Moorov will operate if the individual incidents are of a very different nature, appear relatively benign and/or if the charges relate to more than one complainer. Existing case law suggests that establishing the requisite nexus between incidents may be challenging, for example, where a course of behaviour involves a diverse and varied range of abuse (e.g. sexual, physical, financial, emotional), especially when this takes place over an extended period of time, or where the nature of the behaviour varies from partner to partner.

Part two: Establishing a course of abusive conduct when there is a time gap
Case law has been clear for some time that the nomen iuris of each charge does not need to be identical in order for Moorov to apply. In McMahon v HM Advocate, 17 for example, the High Court confirmed that mutual corroboration could apply between charge 1 (lewd, indecent and libidinous conduct) and charge 4 (assault with intent to rape), with the court stressing that the applicability of the rule depends not on the label attached to the crime but on whether there is underlying similarity of conduct. 18 The modern approach is encapsulated in the following oft-cited statement by Lord Carloway, delivering the opinion of the High Court in MR v HM Advocate: The court today will not proceed upon outdated perceptions . . . but upon its own developing knowledge of sexual and other behaviour and how one type of illegal activity can often be intimately connected with other types of different, but still illegal, acts. Sexual and physical abuse of different kinds perpetrated by one person but occurring within the same family unit, extended or otherwise, is one model of this type . . . 19 In this same case, Lord Carloway emphasised that Moorov can apply regardless of whether there is a difference in the perceived seriousness of the offences libelled. 20 In the appeal case of McAskill v HM Advocate 21 several years later, which arose following the accused's conviction for abuse against three former partners, the High Court confirmed that mutual corroboration could operate between physical and sexual abuse on the basis that the particular charges could 'all be viewed as a course of 16 sustained abuse against a partner in a domestic setting designed, within a context of jealousy, to humiliate and control'. 22 The above decisions appear to offer promise in terms of applying Moorov to the s. 1 offence, which is designed to capture behaviour that varies in perceived seriousness. Nevertheless, there will no doubt be cases where an intimate connection between individual incidents varying in perceived seriousness is not as 'direct and obvious' 23 (see examples further below) and it will be challenging to establish a course of conduct. The challenge will be more pronounced if there is a significant time gap between the different incidents libelled.

The requirement for compelling similarity when there is a lengthy time gap
The case law is now clear that, while there is no maximum time gap beyond which Moorov becomes unavailable, 24 longer time gaps require that there are more compelling similarities between the incidents. The logic underlying this rule is that it is easier to infer a course of conduct when the incidents are more closely related in time. 25 The case of Watson v HM Advocate 26 demonstrates that the doctrine works the other way too, in the sense that a very short time gap between alleged incidents justifies the application of mutual corroboration between offences that differ significantly in terms of seriousness (in Watson, the rape of one woman mutually corroborated the assault of another woman by leg rubbing).
The application of Moorov has thus been denied in several appeal cases with lengthy time gaps on the basis that there were no 'extraordinary', 'special' or 'striking' features linking the different incidents. 27 In the 2017 case of RB v HM Advocate, 28 for example, the appellant was a secondary school teacher who had been convicted of two sexual offences against two pupils, just under 17 years apart, when the pupils were adolescent boys. The second complainer was the nephew of the first, and there were a number of similarities in the acts described by each complainer, including that the appellant: initiated uncomfortable sexual conversations about masturbation; drove to back roads and initiated sexual acts; arranged gym memberships at local gyms; gave massages; and purchased gifts including designer shirts, underwear and drinks e.g. Buckfast wine. 29 Despite these similarities, the High Court allowed the appeal, taking the view that the issue of whether Moorov could apply or not should have been withdrawn from the jury. The court refused to accept the Crown's argument relating to the importance of the family connection between the appellant and the complainers, taking the view that 'the family connection is not a powerful factor' 30 and that the similarities in the evidence were of the type that 'one might expect to find in any two offences of this kind', 31 indicating 'a particular disposition' rather than a systematic course of conduct. 32 While RB suggests that there is a reasonably high bar for establishing 'striking' or 'extraordinary' similarities when there is a lengthy time gap, 33 35 Adam and Daisley was a joint appeal and offers greater clarity in terms of the degree of similarity expected in child sex abuse cases. The High Court referred to cases involving the sexual abuse of children as 'peculiar' and opined that, in such cases, 'there already exists a special, compelling or extraordinary circumstance which will be sufficient for the jury to find a necessary course of conduct established.' 36 It was suggested, however, that the peculiarity of the crime of sexual abuse of children by adults would not be enough if there was 'an exceptionally long gap in time' 37 -in other words that the similarities in the evidence if this were the case would need to be even more special, more compelling or more extraordinary. The earlier decision in RB is inconsistent with this logic. If the starting position is that criminal activity involving the abuse of children satisfies the requirement for compelling or extraordinary similarities, then the claim in RB that the similarities were simply of the kind that 'one might expect to find in any two offences of this kind' 38 becomes untenable (notwithstanding the long timegap). Such inconsistencies in Moorov jurisprudence are nothing new-partly due to its fact-specific nature, over the years the doctrine has evolved gradually in a nonlinear way, with the case law often taking two steps forward before taking one step back.
Before moving on to focus in more detail on the specific types of evidence (or features in the evidence) that can satisfy Moorov's requirement for similarity (or for compelling similarity when there is a long time gap), it is worth highlighting several reasons why the general rule relating to time and similarity may work against the successful application of mutual corroboration in the context of proving the s. 1 offence. First, if the prosecution is seeking to apply Moorov between two or more complainers, then it is highly likely that there will be differences in the precise abuse and acts described by each complainer. Research indicates that the exact dynamics and features of domestic abuse may be unique to individual couples and that specific actions, words or 'gestures, phrases and looks' may only have meaning within the context of a particular relationship (Bettinson and Bishop, 2015: 194). As Tolmie (2018: 54) has explained, each case of domestic abuse may involve: an individualised package of behaviours developed through a process of trial and error for the particular victim by the person who knows her most intimately . . . These behaviours may be subtle and readily understood only by the victim and perpetrator as, for example, when they are designed to exploit fears that are personal to the individual victim.
While Moorov case law is clear that dissimilarities in the evidence do not cancel out similarities, 39 the unique and subjective nature of domestic abuse may mean that there is a reduced likelihood of 'striking' or 'extraordinary' similarities in the evidence compared with other offences. At the same time, however, there is perhaps an increased likelihood of a time gap between incidents when they take place in the context of an intimate partner relationship. This is not only because of the nature of intimate partner relationships, which may of course last for many years, but because victims of domestic abuse routinely delay reporting to the police or do not report at all. 40 (2015). This figure has been described as 'mythical' (Strang et al., 2014: 222), although variable but are often related to the manipulative techniques associated with coercive control, which make victims blame themselves, question their own agency and sanity, make it difficult for them to recognise acts as abusive and, ultimately, to leave the relationship. 41 Experts on coercive control also emphasise that it is misguided to 'negate the impact of the time period between discrete episodes of serious violence-a time period during which the woman may never know when the next incident will occur' as this time period is often experienced as a 'continuing "state of siege"' (Dutton, 1993(Dutton, : 1208. The reduced likelihood of similarities between offences when there is more than one complainer, coupled with the potential for lengthy intervals between reported incidents, means that Moorov's requirement for compelling similarities when there is a lengthy time gap may, in time, give rise to difficulties in certain domestic abuse cases that involve charges under s. 1. This is especially concerning in light of the National Prosecutor for Domestic Abuse's observation that cases involving multiple partners and historical incidents are becoming increasingly common (Hicks, 2014: 12). Although it is this author's understanding that incidents or behaviour that took place prior to the coming into force of the 2018 Act (1st April 2019) could not be prosecuted as part of a 'course of abusive behaviour' under the 2018 Act (due to the general rule against retrospective law-making), it is possible that certain incidents or behaviour could be prosecuted under the 2018 Act and others as, for example, common law assault, stalking or threatening and abusive behaviour (see ss. 38 and 39 of the Criminal Justice and Licencing (Scotland) Act 2010). As noted above, the nomen iuris of each charge does not need to be identical in order for Moorov to apply, meaning that Moorov could be applied to a case involving multiple partners and different charges (some under s. 1 and some not).
However, the 2016 case of ER v HM Advocate 42 demonstrates the problems that can arise in domestic abuse cases involving multiple complainers and a lengthy time gap, with the court refusing to accept the Crown argument that two sets of incidents against three complainers, separated by 24 years, were all underpinned by jealousy, control and possessiveness, and thus a course of domestically abusive conduct. Lord Malcolm sustained a submission of 'no case to answer' from the defence, drawing attention to dissimilarities in the complainers' evidence; the lack of striking, unusual or extraordinary similarities; and evidence from a police officer who had interviewed three of the accused's ex-wives and a former partner, none of whom had reported offending behaviour. The fact that the trial judge concluded that there were no striking similarities despite three complainers speaking to compression of the neck, 43 illustrates the highly subjective nature of judgments about similarity-an issue which is explored in Part four.

Absence of offending during a time gap
In ER, the defence successfully argued that the lack of offending in an intervening period, despite perceived 'opportunity to do so', counted against the existence of a course of conduct. This is a common defence argument in domestic abuse and sexual offence cases. Another noteworthy decision in this regard is the 2019 appeal case of RBA v HM Advocate, 44 in which the accused appealed against his conviction for five sexual offences against two complainers while he was babysitting them. The complainers were related to each other (as aunt and niece) and there was a 13-year time gap between the last alleged incident of abuse against the aunt and the first alleged incident directed towards the niece. The these authors still acknowledge that many will experience years of abuse before reporting. It is also important to acknowledge here that, in a coercive and controlling relationship, there may not be clear, identifiable incidents that are obviously abusive. This point is dealt with later in this article. 41. See generally Stark (2007 High Court accepted the appellant's argument that there would have been a 'continued flow' 45 of children in and out of the appellant's home during the time interval and that the lack of offending despite apparent 'opportunity' counted against the existence of a course of conduct. As Lord Brodie put it: [I]t is to be presumed that an accused has not committed a sexual offence and is not guilty of any act or omission connected with a sexual offence charged in the indictment. There is no room for speculation as to what an accused might have done during any period in respect of which there is no allegation and no proof. Thus, when I refer to a gap in time I mean a period of time when the accused must be taken to have abstained completely from the sort of behaviour libelled in the charges. 46 The court also emphasised that, when seeking to apply Moorov despite a long time gap, the onus is on the prosecution to explain why the doctrine should be applied notwithstanding the length of time between alleged incidents. 47 In this respect, the prosecution may attempt to effectively 'close the time gap' by highlighting possible explanations for a break in a course of conduct. Whereas in child sex abuse cases lack of opportunity may be explained by a generational gap, 48 in domestic abuse cases a break in the course of abusive conduct may be explained by either a period of separation or a spell of time when the accused was not in an intimate partner relationship. As both ER and McAskill illustrate, however, when there is evidence of another relationship without reported abuse the defence will rely upon this to argue against the existence of a course of conduct. In ER, the Crown accepted that it could not rely on lack of opportunity to bridge the time gap as four previous partners interviewed by a Police witness had not reported any abuse during the 24-year gap. 49 In McAskill, on the other hand, one of the issues raised on appeal was the manner in which the trial judge had dealt with evidence from a former partner of the accused. The High Court dismissed the defence argument that the judge had misdirected the jury, and trespassed into their province, when he said that the evidence from a previous partner was of limited value, and questioned the very relevance of the previous relationship because it had been after the incidents libelled. 50 McAskill can be contrasted with ER, where the multiple relationships without reported abuse took place during the intervening period between alleged incidents, and could therefore be more easily perceived by a jury to point against a course of conduct systematically pursued by the accused.
The potential for evidence of intimate relationships without reported abuse to weigh against the existence of a course of conduct for the purposes of Moorov is worthy of comment. A general point to reiterate before doing so, however, is that the successful application of mutual corroboration does not require that there are no breaks in the offending conduct. The words of Lord Sands in Moorov may be referred to here: Opportunity or inclination may be intermittent. A man whose course of conduct is to buy houses, insure them, and burn them down, or to acquire ships, insure them, and scuttle them, or to purport to marry women, defraud and desert them, cannot repeat the offence every month, or even perhaps every six months. 51 With respect to domestic abuse, opportunity or inclination may depend on the dynamics or circumstances of a particular relationship and a non-abusive relationship may simply represent a break in a course of domestically abusive conduct. And, as noted above, these breaks can, in and of themselves, be 45 harmful to the victim when they are experienced as an ongoing 'state of siege' (Dutton, 1993(Dutton, : 1208. While it would clearly be prejudicial to an accused to speculate about behaviour that they may have committed in the absence of evidence, acknowledgment of the reality that there is a reduced probability that such evidence exists in sexual and domestic abuse cases (due to the fact that such offences often take place in private and are notoriously under-reported) is important. Indeed, allowing the absence of offending (despite perceived 'opportunity') to count against the application of mutual corroboration is likely to have a disproportionately negative effect in the very cases that the doctrine was originally designed to assist with.
The decision in Adam and Daisley mitigates the above concern to some extent by confirming that consideration of an absence of conduct in the relevant timeframe forms no part of a legal test, but is a factual matter to be considered by the jury. In this case, counsel for Daisley argued that the appellant would, partly due to his involvement in the scouts, have had the opportunity to engage with other children sexually over the 14-year period during which the offences libelled were alleged to have taken place, and that this had not been addressed by the trial judge despite being raised in submissions. 52 Taken together with the fact that there were 'stark' differences in the evidence and a limited number of charges, this meant that all that could be inferred 'was a propensity to engage in sexual activity with young children' 53 rather than a course of conduct, and therefore that the submission of 'no case to answer' should have been sustained by the trial judge in relation to two of the charges. 54 Refusing the appeals, the High Court confirmed that the significance of an absence of similar conduct was for the jury to consider and 'had no effect on sufficiency'. 55 Applied to the context of domestic abuse, the decision in Adam and Daisley clarifies that the significance of a non-abusive relationship is a matter of fact and degree for the jury to assess when tasked with the ultimate decision on the existence of a course of conduct. As detailed in Part four, jurors should be tasked with this decision unless a trial judge is satisfied that 'on no possible view' could individual episodes be regarded as component parts of a single course of conduct persistently pursued by the accused. 56 Part three: The expansion of the Moorov doctrine As Connelly has previously explained, proving the distinct offence of domestic abuse will in many cases require the underlying similarity to be gleaned from the 'coercive nature and effect of the behaviour and not the actual conduct' (Connelly, 2016: 4). Establishing similarity via the coercive nature and effect of behaviour necessitates a shift away from the traditional application of the Moorov doctrine or, in other words, its further expansion. For the reasons outlined below, this is bound to be contentious, and it is clear that some judges may be more willing to endorse this shift than others. Although (as previously noted) the High Court has been emphatic that mutual corroboration can occur between different charges, 57 several Moorov cases reveal a degree of judicial resistance when it comes to the idea of establishing similarity via motive or mental element. In ER, for instance, Lord Malcolm referred to the idea that a physical assault could support allegations of a sexual nature as 'controversial' and at odds with the 'traditional' view that the 'crimes must be examples of the same crime in any reasonable sense of the term'. 58  coincidence that in both ER and Reilly Crown arguments that separate incidents should be viewed as a campaign of domestic abuse were rejected.
S. 1 of the 2018 Act challenges the traditional view embodied in these two cases as it signals that a wide range of offending behaviour can indeed constitute examples of the same crime, and that there is nothing unreasonable or controversial about seeing incidents designed to control and/or cause fear, alarm and humiliation as inextricably linked. While the policy intent of the legislation is clear, judicial statements in cases like ER and Reilly suggest that the process of carrying this intention through to practice will be far from smooth. Such decisions can perhaps be viewed as an attempt to cling on to the traditional application of the Moorov doctrine following a body of decisions which have expanded the meaning of similarity of conduct. In RBA, Lord Glennie observed that decided cases had stretched the doctrine 'to and beyond their natural breaking point' 60 and questioned the logic of attempting to squeeze incidents into the Moorov 'straitjacket' for the purposes of securing corroboration. 61 The more expansive interpretation of the doctrine required in order to secure convictions under s. 1 will undoubtedly exacerbate long-standing concerns about the erosion of Moorov and, by extension, the corroboration rule itself. Indeed, it is worth making the broader point that the extension of Moorov serves as an argument both for and against the overall abolition of the corroboration requirement: while on the one hand extension supports the view that the requirement is at once not as strong a protection as some might initially assume (Duff, 2012) and overly complex (thus supporting abolition) (Carloway Review: Scottish Government, 2011), on the other hand extension demonstrates that the corroboration requirement is flexible and responsive to social and legal developments (thus supporting retention).
The groundwork for establishing similarity through the nature and effect of behaviour, and therefore even further expansion of the Moorov doctrine, has already been laid by certain appellate level decisions in the High Court. A change in approach has been most evident in child sex abuse cases, where the concept of grooming has been used to establish the requisite similarity required in order for mutual corroboration to apply. In DS v HM Advocate, 62 Lord Brodie observed that the assaults against the two complainers, which were committed after the appellant had built relationships of trust with the complainers' families, 'fell readily into the recognisable pattern of behaviour often described as "grooming"'. 63 As Lord Brodie explained, this concept helped the court to see that the various incidents were planned and systematic in nature, and provided a partial basis for the court's conclusion that there was underlying unity of intent. 64 The court also observed that the time gap of six years and nine months could be explained by the fact that grooming is often a lengthy, drawn-out process. 65 DS therefore indicates that the High Court is ready and willing to use concepts to see connections between, and to develop understanding of, patterns of offending behaviour.
Another emerging theme in Moorov case law, and one that is likely to assist prosecutors seeking to apply mutual corroboration in the context of s. 1, is that the location of the offending, the vulnerability of complainers, and the nature of the relationship between the accused and the complainer/s, are more frequently being identified and accepted as points of similarity. This is illustrated in the 2019 appeal case of McCafferty v HM Advocate, 66 in which the High Court upheld McCafferty's conviction on four sexual charges committed against his former partner and her granddaughter. The appellant's argument that there was no underlying similarity of conduct because some of the charges were committed against a young child in her grandmother's house and elsewhere, while others occurred in the context of a lengthy cohabitation relationship, was rejected, with the court highlighting that 'both were in a domestic setting, 60 with the appellant in a position of trust in relation to the complainers who were both vulnerable in one way or another.' 67 Another case that illustrates the growing relevance of the domestic setting is HM Advocate v MM. 68 This was a Crown appeal following a trial judge's decision to uphold a submission of 'no case to answer' on the basis that there was not sufficient similarity between the charges arising from the frequent sexual abuse of the accused's younger sister, and the subsequent assault, indecent assault and rape of an adult partner 'within the context of a consensual relationship'. 69 The Appeal Court did not agree with the trial judge's assessment that the offending was different in character, highlighting various similarities including that 'both complainers were significantly younger than the respondent, that the relationships both involved the use of violence towards persons with whom the respondent was in a position of some trust, that the alleged offences generally occurred in a domestic context . . . ' 70 The soundness of this approach was questioned in the Scottish Criminal Case Reports commentary to this case: Treating the second set of sexual offences as 'generally domestic', for example, ignores the difference between young siblings and conventional girlfriends, and seems to be another extension of Moorov. Nor is it clear why having a significantly younger girlfriend puts one in a position of some trust. 71 Embodied in this quote is a certain uneasiness about the High Court's understanding of domestic abuse and the dynamics involved in certain relationships. This understanding is evident in other decided cases which, in common with MM, seem to hint at a broader understanding of domestic abuse than that reflected in the 2018 Act's definition of domestic abuse as between partners and ex-partners. 72 In the 2020 case of PGT v HM Advocate, 73 for example, the High Court rejected the appellant's argument that there was only a coincidental family connection between offences committed against the accused's nephew and his ex-wife (and therefore that the offences lacked the requisite similarity) in favour of the Crown's argument that 'the offences occurred in a domestic setting against a background of coercion and control.' 74 The High Court's growing willingness to accept familial connections and the domestic setting as markers of similarity reflects the court's increasingly flexible and contextual approach to establishing similarity for the purposes of Moorov, 75 and indeed to the Moorov doctrine as a whole. The judicial commitment to the modernisation of the Moorov doctrine is also illustrated by the abovementioned case of Taylor, as well as by the 2016 case of Lees v HM Advocate, 76 which confirmed that hearsay evidence (in this case from a deceased complainer) can mutually corroborate another complainer's testimony in court. The above indicates that there is an inclination at High Court level to embrace new and creative ways of seeing connections between offending behaviours, including through employing concepts such as 'grooming'. However, the limit to this inclination has not yet been tested by a reported case where the offending behaviour is markedly different and the nexus between offences is not immediately obvious. In such cases, Moorov's potential to minimise the evidential difficulties with proving the s. 1 offence will hinge on whether there is proper understanding of the policy intent of the 2018 Act (outlined in Part one). It is suggested here that the concept of coercive control could be used in a similar way to the concept of grooming-helping to supply the 'requisite nexus' 77 between incidents, to 'change the complexion' 78 of the prosecution's case, and providing a possible basis for a conclusion that there is unity of intent between incidents. The key issue is whether judges and jurors will be able to see and understand the features of coercive control (including the shared nature and effect of different types of behaviour) to the extent that they are satisfied that there is underlying similarity and therefore a course of conduct. While language associated with coercive control has appeared in Crown arguments and reported decisions, indicating an attempt to embrace the spirit of the 2018 Act, the number of convictions under s. 1 will depend on the depth and scale of judicial understanding of coercive control and, crucially, whether this understanding carries through to juror decision making.
Part four: The challenge of using coercive control to establish unity of intent The problem with the concept of 'similarity' In Reilly (in 2017), the High Court, after referring to the proposed plans to introduce a distinct offence of domestic abuse, seemed to imply that the frequency and regularity of abusive conduct is an indicator of whether a 'campaign of terror such as would characterise domestic abuse' 79 exists. The Crown's argument that there was a systematic pattern of domestic abuse unifying the charges against three complainers was rejected on the basis that two incidents of violent conduct against one of the complainers, PB, were separated by five years. 80 The court reached this conclusion despite there being no indication in the Scottish Government reform proposals-to which the court referred-that the application of a new offence would hinge on different incidents taking place close in time (and indeed there is no such requirement in the 2018 Act), 81 and despite the court accepting that charges involving the two other complainers, RG and DH, were indicative of domestic abuse. The court made the following comment about conduct directed towards DH: we . . . agree that the regular physical assaults which she spoke to were eloquent of what might be called domestic abuse. In her case there were also repeated and varied acts of sexual violence, all interspersed amongst the other violent behaviour to which she was subjected so as to more easily to permit a conclusion that all of the appellant's conduct towards her was of a generally demeaning, coercive and abusive sort. 82 77. Webster, above n. 39 at page 273. 78. AK v HM Advocate [2011] HCJAC 52. 79. Above n. 59 at [34]. 80. Ibid. 81. S. 10 (4) of the Domestic Abuse (Scotland) Act 2018 defines a 'course of behaviour' as involving behaviour 'on at least two occasions'. Nothing in the Act indicates that these occasions must be close together, however some may argue that a reasonable person would not consider incidents separated by five years as 'likely to cause B to suffer physical or psychological harm' (s. 1 (2)(b)). The gap in time in Reilly could point against the existence of a course of conduct for the purposes of Moorov, but at this specific point in the judgment the court is concerned with whether the conduct falls under the umbrella of domestic abuse, rather than whether the time gap counts against the existence of a course of conduct. 82. Reilly (n 59) at [33].
With respect to the evidence of the other complainer, RG, the court commented: 'we can see easily enough that the various physical assaults perpetrated against her are eloquent of conduct which most people would describe as domestic abuse.' 83 Both of these statements indicate that there is an element of judicial discretion involved in assessing what constitutes domestic abuse, 84 and seem to problematically imply both that 'seeing' domestic abuse should be 'easy' and, in the case of the latter statement, that public or common perception of what domestic abuse looks like is relevant to judicial determinations of its existence.
It is also important to question the neutrality and objectivity of assessments of similarity and striking similarity under Moorov. The concept of similarity is inextricably linked with the concept of relevance, which has long been critiqued (by feminist scholars in particular) on the basis that individuals' differing backgrounds, perspectives, education and so on 'interfere' with and de-neutralise determinations of relevance (McColgan, 1996;Ellison, 2010). As McGlynn has demonstrated in her recent critique of the law governing the admission of sexual history evidence in rape trials in England and Wales (McGlynn, 2017), determinations of similarity are plagued with the same problem. These issues are explored by McGlynn in light of the controversial Ched Evans case-R v Evans 85 -in which the Court of Appeal quashed Evans' conviction on the basis that newly emerged evidence of the complainant's sexual activities with third parties would have been admissible at trial because it was 'so similar' that the similarity could not reasonably be explained by coincidence. 86 McGlynn challenges this conclusion, arguing that the identified similarities-the fact the complainant 'had been drinking'; 'instigated certain sexual activity'; 'directed her sexual partner into certain positions' and 'used specific words of encouragement'-were part of 'ordinary and commonplace sexual activity which could reasonably be explained as a coincidence' (McGlynn, 2017: 382) and, further, that the decision increases the likelihood of defence counsel digging through complainants' sexual history looking for similarities (McGlynn, 2017: 384). 87 The subjective nature of assessments of similarity is also highlighted by McGlynn, who rightly suggests that identifying the instigation of sexual activity as a 'similar element' relies upon 'outdated and prejudicial assumptions about women's passivity in sexual activity' (McGlynn, 2017: 382). Such concerns are transferrable to the application of Moorov in s. 1 cases due to the gendered parallels between domestic abuse/coercive control and sexual offences.
Even prior to proposals to create a distinct criminal offence capturing coercive control in Scotland, Scottish policy on tackling Violence Against Women and Girls embraced a gendered understanding of domestic abuse (Burman and Brooks-Hay, 2018: 63-64;Scottish Government, 2018), 88 which, in common with Stark's work, positions domestic abuse as a cause and consequence of systemic gender inequality and power imbalance. Although the new Scottish offence is technically gender-neutral in terms of to whom it applies (both men and women can be prosecuted under s. 1), the notion that domestic abuse both sustains, and is sustained by, patriarchy and male power influenced the drafting of the legislation (for example, the need to maintain a gendered focus was advanced as the primary justification for limiting the offence to partners and ex-partners) (Cairns, 2017: 266 McGlynn also argues that the similarity exception should be removed or, failing that, amended to require a 'demonstrable pattern of highly distinctive and unusual sexual behaviour' (McGlynn, 2017: 390) and notes that New South Wales rejected a similarity exception because 'there would be difficulties in determining when a similarity in the sexual context was sufficiently striking' (McGlynn, 2017: 377). 88. The Scottish Government's approach has attracted some criticism (Dempsey, 2011). gathering to recognise it, and, at the same time, obscuring and minimising its harmful impact' (Bettinson and Bishop, 2018: 8). The authors also express scepticism about the extent to which criminal justice professionals and jurors will be able to appreciate the context in which the offending behaviour occurs if they do not have background or education in domestic violence and coercive control (Bettinson and Bishop, 2018: 9). Tolmie has made similar observations, suggesting that to someone without a gendered understanding of coercive control a perpetrator may 'simply look like an old-fashioned man'-one who expects certain standards in his home and in relation to his children' (Tolmie, 2018: 56). Tolmie goes on to explain that: 'this can be reinforced by women's traditionally devalued status. Women's roles as wives and mothers involve a measure of unpaid servitude, even in otherwise egalitarian relationships, and this can make a victim's oppression difficult to see' (Tolmie, 2018: 56).
The above discussion illustrates why it is so challenging to apply a nebulous concept like 'similarity' in an objective manner, especially in the context of domestic abuse. Seeing similarities, and therefore a course of conduct, will undoubtedly be more challenging if the conduct itself is vulnerable to normalisation and/or having its severity downplayed. In other words, the gendered nature of coercive control likely makes it difficult to recognise certain behaviour as abusive in the first place, let alone to view the behaviour as sufficiently similar to satisfy the requirements of Moorov. Of course, some individuals will be more capable than others of appreciating context and of seeing connections between different incidents-someone who works for a domestic abuse charity and has been trained to recognise the signs of domestic abuse, for example, will be far better equipped than an individual with anti-progressive and misogynist viewpoints. This brings us to the issue of who exactly will be making assessments under Moorov, and of what can be done to maximise their understanding of the dynamics of domestic abuse/ coercive control, both within the confines of the current law and in terms of possible reform.
Part five: The role of judge and jury in determining the applicability of Moorov In recent years there has been a significant growth in the number of domestic abuse cases prosecuted at solemn level in Scotland (Hicks, 2014: 11-12) and it is inevitable that this will continue as more serious cases are brought under the 2018 Act. 89 In solemn cases (which many Moorov cases are) it is the jury, properly directed, who ultimately decide whether separate incidents are sufficiently connected in terms of 'time, place and circumstance', and thus whether a course of conduct exists (Sinclair v HM Advocate). 90 In other words, whether the requisite similarity required by Moorov exists is regarded as a matter of fact and degree. 91 While a case will not go to the jury if the court upholds a defence submission of 'no case to answer' under s. 97 of the Criminal Procedure (Scotland) Act 1995, in Reynolds v HM Advocate 92 Lord Hope stressed that such a submission should only be upheld where 'on no possible view' could it be said that individual incidents were component parts of such a course of conduct. Despite this rule having been confirmed in multiple cases since, including in the abovementioned case of MM, 93 the respective roles of judge and jury in deciding the applicability of Moorov continues to prove contentious, with some even questioning whether it is correct to treat the requirement of a sufficient degree of connection between offences as a matter of fact rather than law. 94 The notion that Reynolds establishes an extremely high bar for upholding a submission of 'no case to answer' has also been challenged, most 89. The Covid-19 pandemic is having a significant impact on jury trials in Scotland. For discussion of this, and the impact of the pandemic on Scottish criminal justice responses to domestic abuse generally, see Cairns and Callander (2020 94. This point is made in the commentary to MM, above n. 68 at 45. In JL, above n. 9, Lady Dorrian opined that: 'it is the function of the judge to assess whether there is sufficiency of evidence. It is for the jury to say whether looking at the evidence as a recently by the appellant's counsel in PGT, 95 who unsuccessfully argued that Reynolds was 'prone to be misunderstood' and that it was wrong to interpret the case as meaning that 'it was only in extreme cases that no case to answer should be upheld '. 96 In theory, the rule in Reynolds should mean that most s. 1 Moorov cases go to the jury and that decisions such as that in ER-in which the case was withdrawn from the jury despite the three complainers speaking to compression of the neck-are increasingly rare. Even where the individual charges do not relate to the same crime or vary in perceived seriousness, the 2018 Act itself and its underlying policy should provide 'the possible view' on which incidents can be regarded as connected. As set out in Part one of this article, 'abusive behaviour' is defined in s. 2 of the 2018 Act as including behaviour that either 'has as its purpose (or among its purposes)' or 'would be considered by a reasonable person' to have one or more of 'the relevant effects'. The relevant effects were included in the legislation to capture psychological abuse or coercive and controlling behaviour that was not previously illegal (Government of the United Kingdom, 2018: point 17) and are defined in terms strongly associated with the concept of coercive control, for example, dependency, isolation, control, regulation, monitoring, deprivation, restriction, humiliation, degradation and punishment. The breadth of s. 2 not only means that a wide and diverse range of behaviour can be prosecuted under the Act, 97 but also that there is considerable scope for establishing connections between isolated incidents for the purposes of Moorov. So long as two or more incidents can, on a 'possible view', be regarded as connected in terms of effect-for example by frightening the complainer/s or restricting their freedom of action-there is a legal basis for the case going to the jury. The incidents can be regarded as 'of the same kind' 98 by virtue of their effect or potential effect on the complainer/s. If there is a lengthy time gap however, owing to the requirement for compelling similarity, it will be more challenging to establish connections via the effect/s of the alleged behaviour (especially if there is more than one complainer). Moreover, there is a risk that the very process of looking at discrete incidents to try and establish connections will entail an episodic focus that proper understanding of coercive control negates against, and that the 2018 Act tries to break away from.
A critical issue, then, becomes how to ensure that judges endorse the spirit of the Act both when assessing 'no case to answer' submissions and directing juries on mutual corroboration. Judicial training on the 2018 Act and domestic abuse will of course be vitally important here and, although it is encouraging that all Scottish judges and sheriffs have already received training on the 2018 Act and the complexities of coercive control, 99 the effectiveness of such training is likely to depend on a variety of factors, including its comprehensiveness and frequency. The knowledge gained in this training will inevitably filter down to judicial directions, which is significant because although jurors make the ultimate decision on the applicability of Moorov in solemn cases, judges should indicate in their directions to the jury if there is evidence before them which would entitle them to draw the conclusion that a course of conduct exists. 100 Despite the Judicial Institute for Scotland's Jury Manual containing a possible form of direction on the application of the Moorov doctrine (Judicial Institute for Scotland, 2019: 15.3-15.4) the adequacy and content of judicial directions on mutual corroboration is an issue that has arisen frequently in recent appeal level decisions. In Adam and Daisley, following a string of defence appeals in which counsel challenged the successful application of Moorov on the grounds of whole; they find it sufficiently compelling to entitle them to conclude that incidents are all component parts in a course of conduct'. 95. PGT, above n. 73. 96. Ibid. at [11]. 97. Indeed, some would argue that the definition of abusive behaviour is too broad. As the Domestic Abuse (Scotland) Bill progressed through the Scottish Parliament, concerns were expressed about the new offence over-criminalising and potentially capturing healthy relationships. 98. Moorov, above n. 3 at page 73. 99. This training was offered by the Judicial Institute for Scotland. Details are available at www.scotland-judiciary.org.uk/25/ 2118/New-domestic-abuse-law-training-for-Scotlands-judiciary (accessed 1 July 2020). 100. Sinclair, above n. 90. misdirection, 101 the High Court confirmed that 'the 'critical direction' in Moorov cases is one which emphasises to the jury that the crimes should be 'so closely linked by their character, the circumstances of their commission and time as to bind them together as parts of a course of criminal conduct systematically (or persistently) pursued by the accused' 102 and that whether there is a need for detailed direction on the requirement for a special or compelling feature depends 'on the circumstances of the case'. 103 This clarification that it is not always necessary for the court to emphasise the need for special or compelling features is potentially significant from the perspective of proving the s. 1 offence due to the reduced likelihood of such features existing in certain domestic abuse cases (see Part two).
Nonetheless, given the unique nature of the 2018 Act and the inherent difficulties with applying Moorov in this context, it is clear that more detailed directions on the nature of domestic/coercive control and its effects would be beneficial from the prosecution's perspective. For the reasons already outlined in this paper, the need for such directions will be most pronounced in cases where the nexus between the charges is not immediately direct and obvious, and jurors are expected to glean similarities by looking at the motive or effect of the behaviour as opposed to the name of the offence or its nature. The inclusion of 'character' in the list of connections that the jury is directed to evaluate is important-arguably even more so than circumstances, time or location-because assessment of the character of the offences connects most easily with consideration of the effects of the behaviour. It is argued here, however, that in many cases more detailed directions on precisely how different incidents can be connected by motive and effects will be required in order to bridge the inevitable gap between the policy underpinning of the legislation and public understanding of the dynamics of domestic abuse. 104 One way of encouraging such directions would be through the introduction of a specific model direction on Moorov and the 2018 Act in the Jury Manual, which could explicitly outline that a course of conduct can be established if crimes are linked by motive (for example, possessiveness, control, punishment, dominance, jealousy) and/or by effect (for example, regulation of activities or behaviour; isolation from friends and family, deprivation of freedom, humiliation). A specific direction is justified by the fact that the successful application of Moorov in this context will, as has already been outlined, in many cases depend on a concerted and unprecedented departure from the traditional idea that mutual corroboration cannot be established by motive and effect. While there is at first sight a 'natural' connection between the Moorov doctrine and the new offence (detailed in Part one), the fact that such a wide and diverse range of different offending behaviour can be libelled together under s. 1 fits uneasily with the traditional view that 'the course of criminal conduct must comprise of crimes which were examples of the same crime'. 105 A carefully drafted model direction, although not compulsory, could help to minimise the potential for appeals on the basis of jury misdirection and would be significantly more straightforward, and less controversial, to introduce than a statutory direction. That said, as the 2018 Act beds in and there is more clarity on how the corroboration rule is operating in this context, an argument may arise for the introduction of a statutory educational direction on domestic abuse that would assist jurors to see connections between different crimes. 106

Conclusion
While proper judicial training and jury directions on domestic abuse/coercive control are necessary, and are likely to have value both in terms of ensuring that Moorov meets its full potential in this context and more generally, it is important to be realistic about the extent to which either can transform deeply engrained attitudes. As Cowan recently emphasised in her analysis of the gap between the 'law on the books' and 'law in action' in Scottish legal responses to sexual violence, the introduction of jury directions, although helpful, 'is not sufficient to shift tenacious stereotypes' (Cowan, 2019: 38). As discussed earlier in this article, gendered stereotypes infiltrate many individuals' understanding of domestic abuse and may make it more difficult for judges and jurors both to see particular behaviour as harmful and to see similarities or connections between different incidents. Training and directions cannot be expected to wash away a host of culturally entrenched expectations and attitudes that have accumulated over the course of a lifetime. While the question of how to fix the genderedness of the criminal law and its key players is beyond the scope of this particular article, it is worth briefly alluding to the potential role that improved judicial diversity in Scotland (specifically through the appointment of more feminist judges) could have in facilitating juror understanding of both gendered concepts like coercive control and the gendered impact of corroboration. 107 Indeed, the significant practical difference that a gendered analysis can make in these areas is illustrated by the recent feminist re-writing of the cases of Ruxton v Lang 108 (concerning the applicability of the defence of necessity in the context of a domestically abusive relationship)  and Smith v Lees 109 (a landmark corroboration case that limited the corroborative value of a complainer's distress in sexual offence cases) (Cairns, 2019) as part of the Scottish Feminist Judgments Project .
In light of the reality that Scotland continues to have a poor record on judicial diversity  and the well-reported limits of the criminal law and justice system in tackling gender-based violence, it is tempting to conclude that the 'gold-standard' Domestic Abuse (Scotland) Act 2018 may in time find itself on the ever-growing list of feminist-inspired law reform that has failed to live up to its full promise. 110 This possibility of unfulfilled potential, and the risk of unintended consequences, has led some commentators in other jurisdictions to express caution over the value of criminalising coercive and controlling behaviours in intimate partner relationships (see, for example, Tolmie, 2018). However, at least in terms of how the Scottish corroboration rule is likely to operate, the doctrinal analysis in this article has revealed that there is some scope for cautious optimism. Despite some inconsistent decisions and Moorov 'fiddles' (Duff, 2012) that appear unsuited to domestic abuse-most notably the requirement for compelling similarity when there is a time gap-a growing body of decisions from the High Court illustrate a shift away from a traditional and strict interpretation of Moorov towards one that is more creative and flexible. The direction of reform, while controversial (and some would argue politically motivated), will inevitably help rather than hinder prosecutions under s. 1 of the 2018 Act.
When it comes to Moorov, however, what one person views as increased flexibility and modernisation, another views as dilution and increased risk of miscarriage of justice. This is a tension that characterised the debate surrounding the abolition of corroboration several years ago, and that feeds into discussions about replacing the Moorov doctrine with a statutory regime governing bad character and similar fact evidence (see, for example, Davidson, 2018). The further stretching of the Moorov doctrine that will be required in order to secure convictions in 'hard' s. 1 cases-i.e. when the offending concerned that such directions are designed to increase conviction rates for sexual offences/domestic abuse and to advance a particular political agenda. For analysis see Callander (2016). 107. For discussion of Scotland's poor record of judicial diversity see Cowan et al. (2019: 25-28 behaviour varies in perceived nature and seriousness and there are multiple complainers over a period of several years-will bring these issues into sharp focus once again. While Scots law has thus far managed to cling to the Moorov doctrine in the face of significant social and legal change, its application in this context may prove to be the last straw that breaks the camel's back.