Theorising sexual harassment and criminalisation in a Swedish context

The #metoo momentum raised not only awareness of the widespread problem of sexual harassment, but also questions as to whether criminal law can provide a proper response.1 Through the #metoo movement, women’s narratives of sexual violence became visible and demands were raised for justice, equality and freedom from sexual harassment. In Sweden, #metoo took shape as 65 mainly occupationally based collective ‘uprisings’ (uppror), but also involved individuals accusing men in powerful positions of engaging in sexual violence. A debate on the impact of #metoo and its


Introduction
This article offers a theoretical approach to criminalisation in relation to sexual harassment, using Sweden as example. The topic is spurred by two separate but interrelated phenomenon: the #metoo movement and an expanding concern that feminism is turning towards criminalisation.
The #metoo momentum raised not only awareness of the widespread problem of sexual harassment, but also questions as to whether criminal law can provide a proper response. 1 Through the #metoo movement, women's narratives of sexual violence became visible and demands were raised for justice, equality and freedom from sexual harassment. In Sweden, #metoo took shape as 65 mainly occupationally based collective 'uprisings' (uppror), but also involved individuals accusing men in powerful positions of engaging in sexual violence. A debate on the impact of #metoo and its contribution to feminist goals in the long run is ongoing. 2 Leaving that discussion aside, it can be observed that #metoo intensified awareness of sexual harassment, and its problem description has been acknowledged by Swedish governmental institutions. The current anti-discrimination legislation, which includes both proactive measures and sanctions against sexual harassment in the workplace, has not fulfilled its promises. Neither have existing criminal law provisions that apply in some instances of intimate intrusions prevented sexual harassment. Quite obviously, the #metoo movement triggered a call for action against sexual harassment.
As the #metoo movement increased in intensity, a process of implementing revised criminal law provisions on rape and sexual abuse -the so-called consent-based rape law -was also taking place. 3 This reform, which had been advocated for almost twenty years, was preceded by several amendments of the sexual offences legislation and should be seen in the context of earlier criminal law measures related to matters of feminist concern. In 1998, the purchase of sexual services became a criminal offence, and a gender-specific domestic violence offence, gross violation of a women's integrity, was introduced. A provision on unlawful persecution (stalking) came into force in 2011, a provision on invasive photography was introduced a couple of years later, and, in 2018, a provision defining the new crime of unlawful violation of integrity was enacted as a response to image-based sexual abuse. Hence, criminalisation has already played an important role in Swedish gender equality politics, and relates to a broader shift towards addressing violence against women in parliamentary gender equality policy that took place during the 1990s. 4 This shift parallels an increasingly repressive crime policy in Sweden, where today, most political parties promote a 'tough on crime' agenda.
In this context, it is perhaps not surprising that Swedish activists have voiced concerns about feminism turning too much to criminalisation. 5 This concern has also been expressed in international research, with Bernstein coining the term 'carceral feminism' to describe how feminist calls for law reform have supported the law and order agen-da in the U.S. 6 Other scholars, however, have opposed this way of framing feminist engagement with law reform. 7 Gotell argues that the carceral feminism critique overstates the influence of feminism on policy and contends that 'the absolute rejection of criminalisation strategies would only intensify the silence around sexual violence as a systemic problem, re-privatising sexual assault and risking the return of impunity for acts of sexual violence ' . 8 Against this backdrop, I suggest that the question of how to define the scope for criminal law interventions regarding sexual harassment needs a more thorough examination from a feminist theory standpoint that takes into consideration the complexities of criminalisation. The following section argues that it is key to start with a theoretically underpinned understanding of sexual harassment, using Liz Kelly's concept of the continuum of sexual violence. After describing this concept, I address some fundamental conflicts that arise when bringing together the continuum concept and criminal law. I emphasise that Kelly's work on sexual violence does not provide an answer as to the extent of criminal law intervention. This, I argue, calls for feminist criminal legal scholarship to take a proactive stance on the issue of criminalisation of sexual harassment. The third section proposes a theoretical approach to criminalisation which captures both formal and substantial aspects of criminal law. The fourth section uses Swedish criminal law on sexual harassment to provide examples of challenges concerning criminalisation of sexual harassment. I use findings from an earlier study on sexual offences in Sweden, 9 combined with new Supreme Court cases, recent legislations and official statistics.

The continuum of sexual violence and criminal law
The feminist framing of sexual harassment has shifted our thinking and language around men's intrusive behaviour against women -e.g. groping, sexist jokes, flashing and intrusive text messages -by connecting this behaviour to gender inequality. Kelly's framing of sexual harassment as part of the continuum of women's experiences of 6 Bernstein, Carceral Politics as Gender Justice? The "Traffic in Women" and Neoliberal Circuits of Crime, Sex, and Rights, 41 Theory and Society (2012)  sexual violence is an important analysis of this interrelation. 10 From interviews with women on their experiences of sexual violence, Kelly created a theoretical tool to understand sexual violence and its function in the gender order of society: the continuum of sexual violence. The theory makes visible 'the complex ways power structures everyday encounters between men and women, and the extent to which intrusion and aggression may be a routine feature of these interactions' . 11 Kelly's work on the continuum of sexual violence has been employed in different fields of research. 12 The continuum concept as it pertains to sexual harassment has been referred to, for example, to study the problem of grey zones, to review criminal law reform, and to study the fear of sexual harassment. 13 A quantitative study in Sweden, Slagen dam, was underpinned by Kelly's analysis and studied the extent of men's violence against women with the specific purpose of gaining knowledge about women's experiences of so-called lenient, but common, violence from men. 14 Research on more specific forms of sexual harassment, such as street harassment, image-based sexual abuse and sexual violence in the digital context, has also used the continuum concept. 15 Its main contribution in these different areas of research is that it allows us to recognise 'the individual impacts, cumulative effects and collective implications of diverse forms of sexual violence in women's lives' . 16 In relation to criminalisation, my point is that a proper criminalisation of sexual harassment needs to consider the harm caused to the individual in terms of the violation of bodily and personal integrity. But more importantly, criminalisation also needs to be motivated by and concerned with the cumulative effects that everyday intimate intrusions cause for women. In addition, criminalisation demands an awareness of the collective impacts engendered by such intrusions, i.e. girls' and women's freedom, safety and security in society. 10 Kelly, Surviving Sexual Violence (Polity Press 1988 The dual meaning of the word 'continuum' is important for the analysis underpinning the continuum concept. First, it means that there is 'a basic common character that underlies many different events' . 17 Second, it means 'a continuous series of elements or events that pass into one another and which cannot be readily distinguished' . In the first sense, 'the basic common character underlying the many different forms of violence is the abuse, intimidation, coercion, intrusion, threat and force men use to control women ' . 18 According to Kelly, the first meaning of the word continuum enables us to discuss sexual violence in a generic sense, and the second meaning allows us to describe the range of abuse, intimidation, coercion, intrusion, threat and force whilst acknowledging that there are no clearly defined analytic categories into which men's behaviour can be placed. An important effect of Kelly's work is that it makes visible the social harm of everyday, routine, intimate intrusions against women. Kelly is clear that the word continuum does not imply a relative seriousness among the different forms of sexual violence (with the exception of violence leading to death). However, Kelly adds, there are forms of sexual violence most women experience in their lives and on multiple occasions. For the purposes of this article, a distinction is made between on the one hand rape, sexual abuse and assault, and on the other sexual harassment. As Kelly points out, there is no clear-cut moment when sexual harassment turns into sexual assault, but the following list largely captures the kind of behaviour this article is concerned with: 'Visual forms of harassment include leering, menacing staring and sexual gestures; verbal forms include whistles, use of innuendo and gossip, sexual joking, propositioning and explicitly threatening remarks; physical forms include unwanted proximity, touching, pinching, patting, deliberately brushing close, grabbing. ' 19 To this list, one might add sexual harassment that takes place in the digital realm. The more common forms of sexual violence are more likely to be defined as acceptable behaviour and thus not to be criminalised.
While this article is concerned with men's everyday intimate intrusions against women, it should be noted that other groups are exposed to similar forms of harassment. The sexism that brings about men's harassment against women is at work in intrusive behaviour directed towards gays, lesbians and bisexuals, as well as nonbinary and transgender people. Sexual harassment is a 'technology of sexism' that reproduces hegemonic masculinities and the gendered order in society which not only women suffer from. 20 According to Boyle, who establishes the notion of 'continuum thinking' , we should think about continuums in the plural, since contexts and connections vary. 21  binary of violence and not violence. My point is that continuum theory can be used to analyse other forms of violence and social power dynamics in relation to criminal law, although in this article, I address the specific problem of men's intrusive behaviour against women.
Kelly's work conceptualises violence independently of established (criminal law) definitions of violence. It hence includes behaviour that does not get much attention from criminal law. 22 Subsequently, a basic tension appears between, on the one hand, the continuum concept and, on the other hand, positivist criminal law regulating violence. The theory of the continuum of sexual violence is grounded in the experiences of women. Thus, violence is defined from a specific gendered position. This contrasts with the gender-neutral language of law and law's claim of being the sovereign, gender-neutral and objective truth-teller. 23 In this way, the feminist concept of the continuum and criminal law are grounded in two very different ontologies. Criminal law is generally concerned with individual instances of violence, which leaves little room to accommodate structural perspectives, such as taking gendered power relations into account. 24 Criminal law also tends to render invisible violence outside the scope of the criminal justice system: the very violence the continuum concept intends to make visible. Law, therefore, plays a crucial role in minimising or denying women's experiences of male sexual violence. 25 Based on two interview studies, Kelly and Radford conclude that women describing intrusions tended to minimise them by saying 'nothing really happened' . 26 Explicitly or implicitly, women connected the intrusion to the risk that they could have been raped, and the fact that rape did not occur led them to say that nothing had happened. Hence, the law, by being there to tell us what is 'real violence' or 'real rape' , influences the way women perceive intrusions that do not amount to a crime under the law.
Another key idea for the continuum analysis is that sexual violence consists of a series of events that pass into one another and cannot be readily distinguished; furthermore, these events cannot be hierarchically ranked according to their seriousness. An inherent logic of criminal law is to grade sexual violence according to its perceived seriousness: rape is considered the most severe crime while verbal harassment (to pick one example) is the least severe. This logic is expressed, for example, in the proportionality principle, which is a guiding principle in sentencing. From a continuum perspective, distinguishing between rape and sexual harassment may make sense to document and unveil the wide range of violence directed against women during their lifetimes. ever, the very notion of the continuum -events that cannot be readily distinguished or hierarchised -stands in contrast with criminal law logic. 27 Given these tensions between the continuum analysis and criminal law, one may feel inclined to abandon the project of criminal law. 28 Kelly and Radford stress that feminist engagement with legal reform has to be aware of how the law functions as a limit to what constitutes sexual violence. 29 Leaving criminal law aside is certainly preferable from a feminist prison abolitionist point of view, which points out the brutality of incarceration and injustices in policing practices. 30 I am reluctant to immediately adopt the concept of carceral feminism, which is mainly grounded in U.S. conditions, into the Swedish context, and there is so far not much support for accusing feminist movements of being the driving force towards a punitive agenda. That said, in recent years, Swedish crime policy has paid increased attention to sexual violations, and feminist calls for justice seem rather easily to become absorbed into an expanding 'tough on crime' discourse. This calls for a feminist legal strategy regarding criminalisation. When arguing for engagement with criminal law, I want to emphasise that this does not necessarily mean promoting more criminalisation, but can mean advocating for laws that are more apt to deal with sexual violence than the current ones do. In addition, we must keep in mind that positivist criminal law is the result of human activity; as such, the principles and logic often presented as the bedrock foundation of criminal law can therefore be amended.
In the Nordic setting, one common approach has been to use the feminist understanding of the continuum of sexual violence as a framework to critically engage with the (in)ability of the criminal legal system to consider women's experiences of violence and, consequently, the failure to properly recognise or punish men's violence against women. 31 Here, the continuum concept serves as a yardstick to evaluate criminal law provisions and criminal justice practice. However, against the background of the current crime policy landscape, I argue that the question of where and how to draw the line for criminal law interventions regarding sexual harassment needs to be more thoroughly examined from a feminist theoretical point of view. A challenge in doing so is that the continuum analysis provides a theoretical framework to bring together routine, everyday intrusions with gender inequality, but does not readily answer the 27 Ibid 51. question of to what extent we should criminalise everyday intrusions. 32 To do that, we also need to conceptualise criminalisation.

Criminalisation
There is a wealth of research on the topic of criminalisation. One strand engages with political philosophy, legal philosophy and criminal law theory, with the aim of defining the boundaries of criminal law and systematising legitimate reasons for criminalisation. 33 Another strand is the field of criminology, which situates crime control in the socioeconomic structure of society, revealing how criminalisation (re)produces social and economic inequalities along class, race and gender lines. 34 As mentioned above, one feminist approach within this field has been to examine the role of criminal law and punitive measures in feminist struggles to end sexual violence. 35 A common denominator between both strands is a critical stance towards penal practices in general and over-criminalisation in particular. In criminal law scholarship, for example, this is expressed in the ultima ratio principle and in creating boundaries for the political use of criminal law, such as the harm principle, or theories of legitimate criminal law interests (Rechtsgüter). 36 As both these strands of research show, there are many reasons for being sceptical about the use of criminalisation as a solution to social problems and gender inequalities. However, the purpose of this article is not to engage in a discussion for or against criminal law as such. Instead, my aim is to shift attention from grand theories on criminalisation to a situated discussion regarding the specific issue at hand: in this case, sexual harassment as part of the continuum of sexual violence against women. This does not imply that we should do away with theory. Instead, as Nicola Lacey in 32 However, Kelly and Radford have asserted that the continuum of sexual violence includes events of violence that should not be regulated by criminal law provisions. It would be 'impossible to legislate against all forms of male behavior which women experience as abusive -that would involve criminalising much of the interaction between men and women' , Kelly and Radford numerous publications has encouraged us to do, 37 we need to pay careful attention to the social and historical context of criminal law theory, provisions and practices, as well as the issue of sexual harassment.
According to Lacey, to understand what criminalisation means in society, we need a 'multi-disciplinary criminalisation research agenda' that is grounded not only in law, criminology or philosophy, but also in history, sociology and political science. 38 The reason for this is a tendency, Lacey argues, in criminalisation research to not make accurate distinctions between formal and substantial criminalisation, for example; or to make normative claims without taking their institutional or political conditions into consideration; or, to take yet another example, to make generalisations without reference to empirical or historical conditions. 39 In short, Lacey argues for a broad conception of criminalisation, including its social, economic, and political implications. That is something more than, or different from, simply asking whether a certain behaviour is or should be in the criminal code. According to Lacey, using the umbrella term criminalisation implies taking into account, among other things: [t]he assumptions, ideologies, ambitions and interests underlying criminal legislation, or the political promises of such legislation; those which inform citizens' decisions to report crime; those informing policing and prosecution decision-making, along with patterns of policing, prosecution and plea-bargaining; the contours of criminal law doctrine and of criminal legislation; the practices of judges and magistrates both in applying criminal law to particular offenders and in sentencing them; the practices of officials in the penal system; even the impact of social attitudes and the inevitable economic costs, the personal ruptures and the knock-on social effects which accompany punishment. 40 I argue that this broad conceptualisation of criminalisation is important from a feminist point of view, where the pressing issue is not restricted to whether there should or should not be a criminal law provision on sexual harassment, but rather is to address the social problem of gender inequality, which includes sexual violence against women. Hence, when considering possible measures, among them criminalisation, we need to comprehend their social effects in a broad sense. That includes, for instance, considering legal regulation other than criminalisation, such as labour law and an- ti-discrimination law, 41 as well as alternatives to formal criminal justice-seeking, such as the emerging practice of digital or viral justice. 42 In addition to the national political context, the role of international conventions, such as the Council of Europe Convention on Violence against Women and Domestic Violence (Istanbul Convention 2011), must be considered. 43 A discussion of criminalisation must also pay attention to victims' understanding of justice, which has been shown to be complex, dynamic and varied. 44 In addition, we need to know more about how criminalisation relates to the way everyday intrusions shape women's lives and the safety work women do to avoid violence, both in the streets and online. 45 And, lastly, we need to be vigilant regarding the risk of laws being enforced in a discriminatory way.
For the purpose of this article, I will focus on one aspect of criminalisation: namely, the challenges in defining criminal law provisions. The next section demonstrates how the scope of criminal law provisions relates to adjudication and what kind of incidents women report to the police, as well as discourses on sexuality. This analysis is underpinned by the distinction between, on the one hand, formal criminalisation (legislation, judicial decisions and international treaties) and, on the other hand, substantial criminalisation (actual implementation of formal norms). 46 This is a way to show, as Antony. Duff has pointed out, that in determining the scope of criminal law -hence, criminalisation -we need to consider adjudication practices, prosecutors' decision-making, the police as agents of criminalisation, and the role of ordinary citizens in the criminal justice system. 47  Conceptualising criminalisation in the manner described above does not entail a rejection of positive criminal law. Rather, my point is to highlight the interrelationship between formal criminalisation and substantial criminalisation. We can think of formal, or positive, criminal law as law in its vertical mode. Margaret Davies uses the images of 'vertical law' and 'horizontal law' , representing two different angles of the law of the nation state. 49 Law in its vertical mode is law thought of as a hierarchical order of norms and principles and a coherent system of norms stemming from an authoritative source. The horizontal dimension of law, on the other hand, thinks of law as plural and considers the social context of law (understood as substantive criminalisation in this article). Davies' main point is to give the vertical aspect of law a position in critical legal thinking, due to the fact that, despite overwhelming arguments showing that law is not an isolated, coherent system, this description of law is also 'true' because lawyers and legal scholars believe in it and act in line with this assumption. 50 To sum up, I argue that we need to conceptualise criminalisation in broad terms when discussing criminalisation of sexual harassment. Hence, it is crucial in an analysis of present and future criminalisation to recognise formal criminalisation (positivist criminal law in its vertical mode) as well as substantial criminalisation (how every element of vertical law has a horizontal aspect; a connection with the social context). Keeping this in mind, I will now turn to the criminalisation of sexual harassment in Sweden.

Sexual harassment and criminalisation in the Swedish context
I have suggested so far that a discussion of the use of criminal law measures in relation to sexual harassment should rest both on a theoretical framing of sexual harassment as part of the continuum of sexual violence and on a broad conceptualisation of criminalisation. In addition, such a discussion requires a review of present criminalisation practices. Against this backdrop, I will now turn to the state of criminalisation of sexual harassment in Sweden. Focusing on law in its vertical mode, my analysis begins with describing the formal criminal law provisions, concluding that the current patchwork of laws does not fit well with the concept of the continuum. Next, I turn to the criminal provision of sexual molestation, which at first seems to offer a criminal law response to sexual harassment. I show, however, that the scope of criminalisation depends on a complex interaction between the formal provision, criminal law principles, adjudication, reports made to the police, and a historically grounded discourse on sexuality.

A. The formal criminal law provisions: A patchwork of laws
There is no specific offence addressing sexual harassment in Sweden. There are, however, several criminal provisions that apply to intimate intrusions. The most obvious one is that of sexual molestation, a catchall provision for deeds that cannot be prosecuted under the heading of more severe sexual offences, such as rape or sexual coercion. 51 It is a sexual offence; hence, the aim of the provision is to protect the sexual integrity and sexual self-determination of individuals. Flashing is explicitly mentioned in the provision. In addition, other types of behaviours (including physical and verbal intrusions) can amount to a crime if the behaviour violates a person's sexual integrity. The scope of the provision thus rests on whether the deed is of such a nature that, from an objective standpoint, it violates the victim's sexual integrity. This objectified assessment implies it is not necessary to prove that the conduct had this impact on the victim and, conversely, the victim's apprehension of the event does not matter. 52 If the deed does not amount to a sexual violation, the criminal law provision on molestation, a crime against liberty and peace, can apply. This provision covers various kinds of behaviour which do not amount to more severe crimes against the person, such as assault, abuse or threat. 53 Its scope of application, however, is limited by the requirement that the deed comprise a severe violation of the victim's peace. Some verbal intrusions, such as verbal insults against another person, usually do not amount to either sexual molestation or molestation. The chapter on defamation offences includes a provision on criminal insult which applies to some verbal intrusions. 54 Likewise, this chapter includes the crime of defamation which can apply to the non-consensual sharing of intimate images. 55 However, and in contrast to other offences, the main rule for defamation offences is that they do not fall within the area of public prosecution. 56 The above-mentioned provisions have been in force since the introduction of the current penal code in 1965. Since 2000, several new criminal law provisions have been introduced targeting specific types of intimate intrusions. One is unlawful persecution, 51 According to Chapter 6, Section 10 of the Swedish Penal Code (Brottsbalken 1962:700), a person who exposes himself or herself in such a manner that the nature thereof evokes nuisance or through words or deed molests a person in such a manner that the nature thereof violates a person's sexual integrity shall be sentenced for sexual molestation to a fine or imprisonment for a maximum of two years. 52 It is similar to the reasonable person test in Anglo-American legal systems: Would a reasonable person find his/her sexual integrity violated by the conduct in question? Compare the UK Sexual Offences Act 2003, Section 78. 53 Chapter 4, Section 7 of the Swedish Penal Code reads as follows: A person who physically molests or subjects another person to disturbing contacts or other reckless conduct, if the deed is suited to severely violate the person's peace, shall be sentenced for molestation to a fine or imprisonment for a maximum of one year. 54 Chapter 5, Section 3, Swedish Penal Code. 55 Chapter 5, Section 1, Swedish Penal Code. 56 Chapter 5, Section 5, Swedish Penal Code. which came into force in 2011 and criminalises repeated violations of the peace: socalled stalking. 57 Another is invasive photography, a criminal law provision on photographic activity constituting invasion of privacy. 58 In 2018, a new crime on unlawful violation of integrity was enacted. This provision makes it illegal to violate a person's privacy through the sharing of images or information about another person's sexuality, health or the fact that the person has been the victim of a crime. 59 It also criminalises the sharing of images of a naked person and images portraying a person in a vulnerable condition.
From the perspective of formal criminalisation, Swedish criminal law provides a comprehensive system of criminal legal regulation of many forms of sexual harassment. The provisions on sexual molestation and molestation criminalise flashing and groping, as well as intrusions via e-mail, phone or social media. The more recent laws on stalking, invasive photography and violation of integrity convey an increasing concern in crime policy with integrity intrusions facilitated by new technology. We can conclude that the Penal Code does not only address 'extreme' forms of sexual violence. In addition, the 2019 report on Sweden by the Group of Experts on Action against Violence against Women and Domestic Violence (GREVIO), the monitoring organ of the Istanbul Convention, found that Swedish criminal law 'gives effect to most of the provisions of the Istanbul Convention' and that the offences named in Article 40 were crimes under Swedish law. 60 While the formal criminalisation of sexual harassment seems to be very comprehensive, I suggest that this patchwork of criminal legal intervention fits less well with the analysis underlying the continuum concept. The provisions were passed at different times in history and are the result of quite diverse crime policy concerns. 61 The result is that the means of intrusion foreground the consequences of such intrusion, that being the social, gendered harm of sexual harassment. Comprehending sexual harassment from a criminal law perspective becomes an exercise in trying to fit these various means of intrusions into existing criminal law categories. This has practical implications: a study on invasive photography shows that the many provisions avail-57 Chapter 4, Section 4b, Swedish Penal Code. 58 Chapter 4, Section 6a, Swedish Penal Code. The law was passed in 2013 and prohibits unlawfully, by technical means, in secrecy, taking a picture of any person who is indoors in a residence or in a lavatory, a dressing room, or other similar space. For example, indecent exposure was criminalised as an offence against morality in the penal code of 1864. able mean additional work in the pre-trial process of deciding under which provision a reported intrusion should be investigated. 62 Regarding the elements that need to be proven for criminal liability, the different provisions vary. For example, the offence of unlawful violation of integrity requires that dissemination of an image is liable to result in serious damage to the person whom the image concerns, while the provision on sexual molestation requires violation of a person's sexual integrity. While there might be reasons to single out certain means of intrusion -as in the case with the taking and dissemination of images -the result is a disruption of continuum thinking and a fragmentation of sexual violence. More importantly, the gendered nature of men's intimate intrusions against women is not acknowledged.
From the continuum perspective, one single criminal law provision on sexual harassment might be preferable to several, at times overlapping, provisions. 63 Such a provision could be modelled in a way that recognises not only the individual impacts but also the cumulative effects that everyday intimate intrusions cause for women, and the collective harm in terms of women's freedom, safety and security in society. It would thereby enhance the possibility to make visible how the different means of intimate intrusions -groping, verbal intrusions, flashing, etc. -share the common character of controlling women and serving as technologies of sexism. A single provision on sexual harassment that considers the several dimensions of individual instances of intrusive behaviour is potentially a more accurate way to distinguish behaviour that deserves to be punished from conduct that is merely inappropriate. Arguing for a new criminal law provision on sexual harassment should therefore not be perceived as contributing to an excessive use of criminal law. Further, a review of current provisions may well result in arguments in favour of the decriminalisation of provisions that are obsolete in aims and scope. A provision on sexual harassment could render redundant both old provisions, such as criminal insult, and recently introduced crimes, such as invasive photography. Although decriminalisation is far less popular in crime policy, current criminal law provisions are in fact open for negotiation.
If a single provision on sexual harassment is more suitable to the concept of the continuum than a patchwork of laws governing different everyday intrusions, it remains to define such a crime. I will now draw attention to some of the difficulties in doing so, taking the Swedish criminal law provision on sexual molestation as an example.

B. The criminal law provision on sexual molestation: The role of adjudication and police reporting
The wording of the provision on sexual molestation suggests that it could encompass most forms of intrusive sexist behaviour. What is required is that the offender, through words or deed, violates another person's sexual integrity with intent. From the perspective of the continuum concept, such an open-ended definition of the actus reus might be preferable. This vague definition, however, has left the task of defining the crime to the discretion of the courts. In court practice, criminal liability for sexual molestation can be excluded for various reasons. 64 Leaving evidentiary questions aside, the courts can refer to the general conditions for criminal liability (e.g., lack of intent) and to criminal law principles (e.g., the principle of legality). Furthermore, the provision on sexual molestation may not apply due to conditions specific to the category of sexual offences (i.e., perceived consent from the molested person, or the deed not being directed towards a specific person or group of persons). The courts, moreover, can refer to requirements and examples provided in preparatory work concerning sexual molestation (the deed must be of a clearly sexual nature or conducted for the purpose of sexual gratification). I mention these examples to point out that the contours of the provision on sexual molestation cannot be easily drawn. Not surprisingly, this conclusion fits with the theories on criminalisation discussed in Section 4; it is not the case just for the crime of sexual molestation. It is, however, a concrete example of how the boundaries of criminal law provisions depend upon many different features of the criminal legal system.
When analysing the role of adjudication in criminalisation, police reporting practices must be considered. One aspect of including reports made to the police in a study of criminalisation is analysing the extent to which formal criminalisation corresponds with substantial criminalisation. Is sexual molestation a de facto matter for the crim-  Among other findings, the study shows an increase in the share of physical sexual molestation and in the share of visual sexual molestation, but a decrease in the share of verbal sexual molestation. 69 In addition, the number of reports of physical sexual molestation increased. 70 If this category is subdivided into a) touching of genitals/ breasts and b) touching of other body parts, the bulk of the increase came in the latter category. 71 The study shows that reports made to the police have increased, and that the increase has taken place across multiple categories of intrusion.
Another aspect of criminalisation and practices of reporting to the police is that the vertical scope of criminalisation can be challenged horizontally, as court practice interacts with police reporting and prosecutor decision-making. An increase in reporting to the police can lead to courts having to rule on cases that had not previously entered the criminal justice system. Attrition -the process by which cases are discontinued and thus fail to reach trial -as well as decision-making by public prosecutors need to be explored further to address this issue. However, I would argue that the increasing number, in recent years, of Supreme Court cases on sexual molestation should be seen against this backdrop. Since the passage of the current Penal Code in 1965, published cases on sexual molestation have been rare -there was one case in 1988 and two cases in the 1990s. 72 But recent years, the Supreme Court has heard four cases concerning sexual molestation. In one case, the Court found the provision on sexual molestation to be applicable to a situation where a man had repeatedly, in a public place, asked a young woman to have sex with him in exchange for money. 73 The provision was also found applicable to so-called 'up-skirting' , in a case where a man in an escalator had put his cell phone under the skirt of a woman to take a picture of her genital area. 74 Two other cases concerned the unwanted touching of body parts other than genitals and breasts. In both cases, the Supreme Court found that the provision on sexual molestation did not apply. In the first case, the victim was a 15-year-old girl and the defendant was her boyfriend's father. 75  which the deed took place (among which was that the defendant subsequently told the victim he had feelings for her), it could not conclude that the deed was of the sexual nature required for criminal responsibility. In the second case, a man was charged with sexually molesting a woman by touching the inside of her thigh. 76 The defendant was, at the time, the director of a municipal department where the victim worked as a trainee. The alleged offence took place at an after-work event. The Court concluded, after noting that the touching was brief and that the woman was wearing jeans, that the defendant's behaviour was both inappropriate and unwelcome but not so clearly sexual in nature as to fall within the scope of criminalisation. I will not analyse these cases in detail here, but only make the point that the 2018 Supreme Court cases can be perceived as a response to the increase in police reports of unwanted touching of body parts other than genitals/breasts.
To conclude, the significance of court discretion in criminalisation makes it difficult both to pin down and to change the scope of the criminal law provision on sexual molestation. Court practice, however, follows what has been previously been done in the criminal justice system: prior cases as well as governmental bills and reports and law commentaries. As the Supreme Court cases described here show, the requirement that the deed should be of a sexual nature is often referred to in court practice and in preparatory works for drawing the boundary for criminal liability. To comprehend the scope of the provision, therefore, we need to understand what the requirement of 'sexual' signifies. One way is to explore the historical context of sexual offences.

C. The criminal law provision on sexual molestation: Understanding sexual integrity in a historical context
The provision on sexual molestation covers deeds that violate the sexual integrity of another person. Following Supreme Court practice, the deed must have a clearly sexual nature. In earlier research, I found that this concept of sexual integrity has more to do with sex, as in erotic sexuality, than with gender, as in Kelly's definition of sexual violence. 77 This, I argue, is due to the historical development of sexual offences. The genealogy of sexual offences in Swedish criminal law is usually described in terms of a narrative in which the primary good (Rechtsgut) is supposed to have shifted from public morality in the nineteenth century to individual integrity in the mid-twentieth century. According to this narrative, the law on sexual offences has gradually improved since the Penal Code of 1864, so that the legal protection of sexual integrity has become increasingly comprehensive. This is true in some respects: of rape and the provisions governing sexual exploitation of children, for example have expanded.
I was intrigued, however, by the idea of sexual integrity as a guiding principle for criminalisation. I argue that when sexual integrity became a protected legal interest (a Rechtsgut) in the mid-twentieth century, sexual offences also came to be understood through the lens of erotic sexuality. Responsibility for sexual molestation, for example, was often described as requiring the offender to have acted for the purpose of sexual gratification. During this period of time, the discourse of crime policy constructed sexuality as something positive and worthy of protection, and something that belongs to the subject's inner personality. Sexual offences became understandable only in relation to the lustful, consensual and erotic sexuality. A strict distinction was set up between violence, on the one hand, and sexual offences, on the other. According to this dominant construction, sexual offences are understood as a matter of sexuality, where gender is absent and the crime policy problem is one of individuals rather than of structural gendered inequality. 78 This construction of sexuality in criminal law discourse derives from, and reproduces, how the boundaries of sexual molestation are drawn in court practice. Because of the way the 'violation of sexual integrity' element is interpreted, behaviours need to have a clear sexual meaning or character, namely a connection to erotic sexuality, to amount to a crime. The crime policy discourse on sexuality affects how the legally relevant context is defined in court practice. The defendants' behaviour in a narrow sense, and their motives, become central, and although harms to the complainant might be addressed, these equally focus on the individual level.
Summing up, from the perspective of the continuum concept, all of the deeds prosecuted in the four Supreme Court cases -verbal intimidation, up-skirting and unwelcome touching -are examples of everyday, routine intrusions on the continuum of sexual violence. The Supreme Court considered the first two as criminal offences, but not the latter two, mainly because the behaviour in question lacked a sexual quality. Drawing on feminist research on sexual violence, such as Kelly's, we could argue for criminalising this kind of sexist behaviour. We know that in relation to the scale of violence that criminal law deals, an individual instance might seem harmless, but the cumulative and collective effects are harmful for society. A single provision criminalising sexual harassment -such as the Swedish one on sexual molestation -has the potential for more accurately allocating criminal responsibility, given that sexual in-78 However, when analysing legislative processes and court practice, I also found an alternative construction of sexual integrity according to which sexual offenses constitute gendered violence, sexual offenses are described as men's violence against women and sexual integrity is a gender equality issue. Compare Andersson, who in a discourse analysis of court practice on sexual offences found that the subject of criminal protection is constructed as feminine, with a passive sexuality and a body that is open and accessible. Andersson, Hans (ord) eller hennes? En könsteoretisk analys av straffrättsligt skydd mot sexuella övergrepp (Bokbox 2004). tegrity is understood as women's freedom from everyday routine intrusions, and that the scope of criminalisation does not depend on the sexual nature of the deed, but instead on whether the behaviour is sexist and reproduces gender inequality. That said, I am not convinced that every sexist intrusion should be subject to criminalisation. I have shown that where and how the line for criminal law intervention is drawn is a complex process which depends not only on adjudication, prosecutors' decisions and police reporting practice, but also on the historical context of the regulation of sexual offences. Any attempts to design a criminal provision regarding sexual harassment from a feminist legal theory standpoint will need to consider this complexity.

Conclusion
In Sweden and globally, women have been reporting sexual violence for decades. However, the #metoo movement marks a substantial shift, not only for the urgency, immediacy and global reach of the digital campaign, but also in the public reception of women's reports of sexual harassment. #Metoo has triggered a call for action against sexual harassment, raising the question of to what extent criminalisation can and should serve as a proper response. This question has several dimensions, one of which is the role of criminalisation in feminist activism and the extent to which punitive measures are compatible with the broader aim of social justice. While recognising the importance of that discussion, I contend that it should not be limited to a choice between either criminalisation or no criminal law measures at all. To a large extent, everyday intimate intrusions are already formally criminalised in Sweden and increasingly reported to the police.
I therefore suggest that feminist research and activism should engage in a critical conversation with criminal law that takes into consideration the injustices that accompany a more punitive crime policy climate. 79 This conversation needs to examine how to define the scope for criminal law interventions regarding sexual harassment from a feminist legal theory standpoint, and this article is one contribution to such a project. I have highlighted the fact that although the continuum analysis provides a theoretical framework for bringing together everyday intimate intrusions and gender inequality, it does not provide much guidance on the extent to which such intrusions should be criminalised. In addition, I have made visible some of the challenges that come with designing a criminal law provision that would consider both individual effects as well as the cumulative and collective impacts of men's intimate intrusions against women. These challenges should not prevent us from engaging with criminal law. As pointed out in this article, such an engagement does not necessarily entail more criminalisa- 79 For detailed discussions on strategy, see e.g. Gotell 2015, and Terwiel 2020.