1On 7 July 2007 Australia’s largest circulation daily newspaper, Melbourne’s Herald-Sun, ran a front page banner headline Village of the Damned: Keeping an Eye on Our Worst Sex Monsters. Accompanying it was a photo of a grey haired and balding man in a rumpled suit jacket, reading glasses tucked into his woollen vest, presenting to readers a figure who, for any other purposes, might be a kind of every-man uncle or grandfather. But the man was in fact a rather well known sex offender who, at the conclusion of his custodial sentence, had been made subject to one of the state’s newly minted Continuing Supervision Orders, the terms of which may last for up to 15 years and include restrictions upon movement, residence, acquaintance, access to the internet and more. The Village of the Damned to which the newspaper article referred was in fact the residence of 10 such individuals, lying inside the razor wire of a regional prison, the only location that could be found that would both secure their safety from community vigilantes and satisfy the draconian oversight requirements of a law the Victorian Supreme Court had described as making otherwise free citizens ‘a prisoner in all but name’ (TSL v Secretary to the Department of Justice, 2006). But rather than being unique or remarkable, this law in fact forms part of a new governmental logic and strategy. It dovetails neatly with a series of other new restrictions and pre-emptive risk containment measures directed at suspect characters, most notably including terrorists and illegal migrants, and which materialises a number of formerly inchoate tendencies to extend punitive practices into the realm of civil threat or disorder.
- 1 Australia is divided into six states and two territories, overseen by a Commonwealth government. C (...)
2This new approach to sex offenders in Australia began in June 2003 when Queensland’s Dangerous Prisoners (Sexual Offenders) Act came into force.1 There now exists a small literature on the Act and the case of Fardon v Attorney General for the State of Queensland , in which the constitutionality of Robert Fardon’s indefinite preventive confinement under the Act was upheld by the High Court (see for example, Gray 2005, Keyzer and Blay, 2006; McSherry and Keyzer, 2009; Mercado and Ogloff, 2006; Roos, 2005). In 2005 the state of South Australia amended its Criminal Law (Sentencing) Act 1988 to provide similar measures. And in 2006 both Western Australia and New South Wales responded, promulgating legislation that provided for post-sentence detention and/or extended supervision orders under the Crimes (Serious Sexual Offenders) Act (NSW) 2006 and the Dangerous Sexual Offenders Act (WA) 2006. The state of Victoria introduced continued supervision arrangements (for up to 15 years, renewable) in 2005 and in January 2010 these were supplemented by measures for continued detention introduced in the Serious Sex Offenders (Detention and Supervision) Act 2009. Each of these schemes provides for long periods of post-sentence confinement – ranging from indefinite detention in South Australia, Queensland and Western Australia to five-year (extendable) detentions in New South Wales and three-year (extendable) periods in the Victorian scheme. These confinement schemes may also be buttressed by arrangements for the long-term supervision and monitoring of offenders in the community, such as the Victorian provisions described above, and through legislation providing for the lifetime registration of ex-sex offenders.
3While the post-sentence character of these measures is new, Australia along with other common law jurisdictions has a long history of sentencing offenders to preventive detention. Notwithstanding this, the popularity of preventive measures has risen and fallen over time. At the turn of the 20th century all Australian states adopted indeterminate sentencing laws modelled upon the Habitual Criminals Act 1905 (NSW). Around mid-century anxieties about the psychopathic offender became prominent and were reflected in the emergence of defective offender statutes and revisions to habitual offender legislation, traces of which remain today such as in Queensland’s Criminal Law Amendment Act 1945 and NSW’s Habitual Criminals Act 1957. Yet by the 1980s most of these had fallen into disuse or irrelevance. Indeed as early as 1968 Victoria’s Director of Prisons was able to describe the state’s indefinite detention provisions as a dead letter, with only one offender having been sentenced under s 537 of the Crimes Act 1958 (Vic) in the whole preceding decade (Daunton-Fear, 1969). Beginning in the 1990s these schemes again began to find favour, but it was still possible a decade later for Arie Freiberg, a long time observer of sentencing practices, to observe that, overall, Australian ‘judges … were not prepared to countenance legislative alternatives which were regarded by them as being more dangerous than the dangerous they sought to govern.’ (2000: 58). Indeed, as an illustration of this, Victoria’s own revitalised preventive detention sentence has been applied to just four offenders since its reintroduction in 1993: not much more than the once-a-decade use observed during the 1950s and 1960s.
4For critical criminologists the question at hand is how best to make sense of these new forms of involuntary post-sentence preventive control. On the one hand, the history of western modernity offers an extensive catalogue of ways and means by which ‘suspect’ populations have been surveilled, controlled and managed. Many of these penal rationalities and governmental techniques were established and trialled in Europe’s colonies (Brown, 2004; Pierce and Rao, 2006) and more recently such colonial strategies have been reprised both in domestic (Brown, 2005) and international settings (Gregory, 2004), illustrating the enormous continuity that characterises modern penality (David Brown, 2005; Sim, 2009). The modalities of penal technique and bureaucratic justice now directed at suspect groups such as immigrants are also not new, as Clifford Rosenberg (2006), for example, illustrates in his account of police surveillance and control of immigrants in Paris between the two world wars. On one view then, contemporary developments might be seen as merely more of the same. Yet there appear, prima facie at least, to be a number of important differences between former times and the present.
5The second section of this paper will therefore consider the question of how best present developments in involuntary detention might be framed in theoretical terms. Is the advent of post-sentence preventive detention best thought of through a lens that puts, for example, risk at the theoretical centre? Or is risk perhaps better understood as an element of the idea of dangerousness, so that the latter might offer a more productive alternative. Does the removal of an individual from society abrogate fundamental rights, and if so would not rights and liberties be the best way to approach this new phenomenon? This section will in fact conclude that a theoretical approach developing the idea of a jurisprudence of risk – that point where the polyvalent tropes of risk, freedom and security meet the practices of law and government – offers the most fruitful way of understanding contemporary realities in involuntary post-sentence detention. Section three will examine the emerging jurisprudence of risk in Australia through attention to a body of appellate case law. The focus will be the case of Director of Public Prosecutions (WA) v GTR , a case which draws upon and condenses much of the extant Australian case law on indefinite involuntary detention of sex offenders. What will be revealed here is the abrogation of normal standards of justice that apply when courts are faced with questions of imprisonment for punishment. Instead, a new hybrid domain is established, revealed in shifting standards aimed at accommodation of governmental objectives of securing continued detention despite its repugnance to set principles. Finally, section four of the paper contemplates these findings in their broader theoretical context, suggesting how the suspension of legal normativity evidenced in recent Australian case law can be understood within a discourse of liberal governmentality, security and the jurisprudence of risk in the context of danger and emergency.
6Analytically, one way to understand post-sentence involuntary preventive detention is to situate it within the rapidly expanding field of security in the post-9/11 world (Sorkin, 2008). This is particularly so in Australia, where constitutional protections against involuntary detention have been consistently and quite radically eroded in the wake of September 11 events (see McHugh, 2006). This has occurred not only in the emergence of detention and control-order regimes for ‘terror suspects’ (Bennett, 2007), but also in the extension of a now widely recognised precautionary logic (Goldsmith, 2008) to asylum seekers (Grewcock, 2010), bikie gangs (Lynch, 2009) and sex offenders. The undermining of traditional legal and constitutional protections has occurred, moreover, in a way and at a pace that is quite out of step with developments in Europe. Whereas European legal institutions, such as the European Court of Human Rights, have in recent times sought to reinforce principles of justice and liberty and protections against an overbearing state (see e.g., M v Germany ), Australian courts have moved strongly in the opposite direction, meaning that many in Europe might be surprised by the size of the gulf that now separates the two legal regimes. For all of these reasons, attention to notions of security, liberty and right would seem to be crucial ingredients in any analysis of involuntary preventive detention regimes for sex offenders.
7Before pursuing this, however, it must be recognised that all these preventive responses – whether they seek protections from terrorists, bikie gangs or sex offenders – are crucially linked to the wider and pervasive influence of risk in modern western societies (see generally, O’Malley, 2004). Critical understandings of risk as a phenomenon are by no means monolithic and several strands to the risk paradigm exist. For present purposes, we may observe that risk thinking tends to achieve its effects through a double movement. On the one hand the notion of risk works to render the broader social space a field of perceivable and calculable risks: of environmental damage, or industrial process failures, of negative health or welfare outcomes, or crime events (Beck, 1992; Douglas, 1992). Risk in this sense is expansive. It is a heuristic that changes everything placed before it, reframing social and technical phenomena alike into nodal points on a society-wide grid of risk coordinates. But the notion of risk also works in the opposite fashion, tending to condense around specific sites of social anxiety that then become locations of intense public debate and regulation as well as appearing in the public mind as archetypal scenes and forms of danger: current examples include those cited earlier, such as the terrorist, the illegal immigrant/asylum seeker, the bikie or the sex offender (Brown and Pratt, 2000; Castel, 1991; Simon, 1998; Spencer, 2009). Since we are principally concerned here with the latter, the figure of the sex offender as both a site of danger and a crystallization of dangerousness, it will be worthwhile examining this formation in some more theoretical detail.
8In discussing the extension of a precautionary security logic to sex offenders, Hebenton and Seddon (2009: 355) wonder aloud ‘why it is that, across diverse cultural contexts … (United States, United Kingdom, Australia), the sexually violent offender has become the exemplar par excellence of this “emergent response”’ of precaution. Of course sex offenders have always generated a certain amount of attention and anxiety, as reflected for example in the sexual psychopath laws of 1950s America that were directed against gay men (Miller, 2002). But this focus as also ebbed and flowed. In the cycle of dangerous offender anxiety that swept English speaking common law jurisdictions in the 1980s it was violence and the unpredictability of mentally disordered offenders that captured attention (Floud and Young, 1981; Gunn and Farrington, 1982). So the return of sex offenders to the heart of cultural fears about risk and danger is important to understand.
9In common with other criminologists, however, Hebenton and Seddon have been at a loss to say why, specifically, sex offenders rather than any other offender type should so perfectly personify and crystallize notions of risk and danger across such diverse cultural sites (see for example, Lynch, 2002; Simon, 1998, Spencer, 2009). Perhaps as a result of this difficulty, most criminological work on the sex offender has tended toward a descriptive model of theorization. Mona Lynch, for instance, draws upon Mary Douglas’s early anthropological work on danger and taboo, to situate representations of the sex offender in a field of ‘emotional expressions of disgust, fear of contagion, and pollution avoidance’ (2002: 532). Dale Spencer does much the same, but replaces Lynch’s expressive emotionalism argument with a not dissimilar proposal, following Giorgio Agamben, that the sex offender may understood through the analytic of homo sacer, ‘the life deemed dirty, impure or accursed … which no one can touch without dirtying oneself’ (2009: 224). In each case, these writers seek to account for the positioning and treatment of the sex offender in the social field (as reflective of disgust, or exceptional status). But they do not say why sex offenders should be so uniquely positioned: why, for instance, child sex offenders and not child traffickers?
10Drawing upon two relatively separate strands of work by Michel Foucault, I have recently suggested how the persistence and vitality of fears of the sex offender can be understood (Brown, 2010). The two strands upon which I draw are Foucault’s discussion of the monster, appearing in the Abnormal lectures given at the College de France in 1974-75 (Foucault, 2003), and his work on governmentality and security that appeared in his 1977-78 lectures titled Security, Territory, Population (Foucault, 2007). Whereas Foucault had earlier been concerned with the way sovereign and disciplinary power together pursued a dream of ‘homogeneous, continuous, and exhaustive’ control through domination (2007: 66), in what he would later describe as a governmentality of police, his work in Abnormal reflected a new interest in the natural processes to be found in nature itself. While it is common even among contemporary criminologists to use the epithet ‘monster’ to describe sex offenders, positioning them as something apart from us (eg., O’Malley, 2000; Simon, 1998), Foucault recognised in the monster something quite the reverse. For him, the monster was the ‘spontaneous, brutal, but consequently natural form of the unnatural’ (Foucault 2003: 56). Looking at the monster, he argued, we find ‘the magnifying model, the form of every possible little irregularity exhibited by the games of nature … the major model of every little deviation … [and] the principle of intelligibility of all the forms that circulate as the small change of abnormality’ (Ibid). Within the liberal governmentality of contemporary society, mechanisms of security are invoked to control the dangers inherent in such nature taking its course: to manage the risks inherent in freedom itself, which is after all the engine of the liberal form of government. (Foucault, 2007). If we take the analysis of monsters given in Abnormal and integrate it with Foucault’s work on governmentality, it is possible understand the monster – of whom I will suggest the sex offender is a kind of modern incarnation – as a major model of all the deviations and errant potentials, and therefore all the risks and dangers, to be found in male sexual nature itself. In this context male sexual nature becomes a target of governance as liberalism ‘turns into a mechanism continually having to arbitrate between the freedom and security of individuals by reference to the notion of danger’ (Foucault 2008: 66).
11The liberal rationality at play here sheds the old binaries of good and evil, of normal and abnormal, replacing them with regulatory mechanisms of security that analyse and transect the social field in terms of degrees of risk. In this process, the sex offender becomes connected with ordinary men across a seamless and continuous field of sexual regulation. This field is bounded at one end by the forms of petty regulation that require adult men to, for example, present identification when collecting a child from school, and forms of subjectification that shape men’s conduct around children, such as in male child care workers’ inclination never to be left alone with a child (The Times 2007). This wide regulatory domain is marked at its opposite end by the extensive and illiberal measures of precautionary and preventive security to be found in post-sentence preventive detention schemes such as those now proliferating across Australia. What explains the resilience and vitality of the sex offender as an image of danger and as an object of illiberal security mechanisms should therefore be clear. It is that the target of these measures is only partly offending behaviour in its own right, in the sense of its criminality, for the main target of this regulation is in fact tendencies that are natural or immanent in the population as a whole. In other words, the target of precautionary, preventive detention is not just the sex offender but indeed human nature itself.
12If the foregoing allows us to understand to some extent why sex offenders and not, for example, people traffickers, bank robbers or methamphetamine cooks, crystallise so effectively cultural fears and anxieties and become targets of illiberal mechanisms of governmental security, it is less effective in allowing us to understand how such mechanisms of security work. To understand security in practice we need to move from the meta-theoretical domain of liberal governmentality to a theory of practice. Specifically, if security is as Foucault suggests ‘the fundamental objective of governmentality’ (2007: 353), how will it be managed, partialled out or allocated? In broad terms, Foucault does answer this: in a liberal governmentality based upon freedom, security will be needed to ‘ensure that the mechanism of interests’ which freedom allows into play ‘does not give rise to individual or collective dangers’ (Foucault 2008: 65). But security will always involve some degree of balance, since those very mechanisms designed to contain risk and danger may ‘produce destructive effects which prevail over the very freedom they are supposed to produce’ (2008: 69), creating what Foucault termed ‘liberogenic’ effects (Ibid.).
13To understand security in its operationalisation and practice, to obtain a finer grain of analysis, we need to move beyond Foucault. Many scholars have taken up the notion of balance inherent in his characterisation of the ‘liberogenic’ effects of security. Mark Neocleous (2007), for example, challenges Foucault’s characterisation of security as a mechanism inherent to liberal freedom and disavows the ‘myth’ of balance. He proposes, contra Foucault, that ‘liberalism’s central thematic is not liberty, but security’ (p. 133), arguing thus that any proposed accommodation between freedom and security is but a chimera. Lucia Zedner, whose work is concerned with the tension security invokes between different conceptions of rights, is another person who has done much to develop the analysis of security’s relationship to freedom (Zedner, 2003a, 2003b, 2005, 2007a, 2007b). One the one hand, she suggests, there is the widely held assumption that the key justification and purpose of security is the protection of freedom, yet the means by which this often ill defined notion of freedom is achieved often works to undermine those very liberty rights that constitute political freedom. Looking to the intelligibility of such practices, she cites Michael Ignatieff’s observation that such undermining reflects what he terms a ‘moral temptation’ in our society to see liberty itself as divisible, such that the liberty of some may be secured by extinguishing or significantly constraining the liberty rights of others (Zedner, 2005: 524). Part of what makes such a strategy so tempting is that the costs of public safety achieved in this way tend to fall on just a small minority who are felt to be, as it were, dispensable, not only in the utilitarian sense of their small number but also through their being what Nils Christie and Loic Wacquant have termed in the penal context ‘suitable enemies’ (Christie, 1986; Wacquant, 1999). The danger associated with this sort of approach to securing social goods forms a conceptual thread that runs throughout Zedner’s work. It is not intended here to reprise her arguments in detail, but rather to draw out a few key observations that will provide a springboard into the present paper, the aim of which is to consider the emergence of one apparatus of security in Australia.
14Four observations will help set the scene. First, despite its frequently paradoxical (or incoherent) effects, and despite the many difficulties with the notion of ‘balance’ that is used in its justification, the recruitment of security measures is ultimately tied to the desire to preserve liberty and freedom (Zedner, 2003a, 2005). The observed failure of security measures to do so, or their inherent tendency to trample the very goods they are supposed to uphold – Foucault’s (2007) notion of their ‘liberogenic’ effects – suggests therefore that it is quite possible to have too much security (Zedner, 2007b), but this still leaves unanswered the question of how security might best be measured out. Second, the failure of security mechanisms, and indeed of the demand for them, to self-regulate naturally means we should look for analogous situations that would supply principles for the justification and distribution of security. Zedner argues in favour of ‘harvesting … analogous principles’ (2003a: 176) from criminal justice, where experience with the high-stakes associated with punishment would provide an approach to security that offers appropriate protections to individual rights. Thus, a principled approach based upon the procedural and structural safeguards common to criminal justice is suggested by Zedner as an appropriate corrective to rhetorical or politically flimsy arguments about the appropriate balancing of individual liberty rights with public safety. Third, as noted previously, special attention is needed to the way in which risk assessment and threat profiling has created new classes of dangerous individuals whose behaviour is rendered, at least notionally, scientifically predictable. It is precisely this rational scientific measurement of risk, argues Zedner, that ‘that furnishes the ostensible moral basis for designating dangerous populations whose rights to due process are overridden solely because they belong to a particular group’ (2007b: 25). Finally, it must be recognised that the discussion of liberty rights is always twofold: liberty involves both a freedom to act (typically termed positive liberty) and a freedom from governmental restriction (negative liberty) (see also, Foucault’s discussion of this fundamental distinction, 2008: 39-45). When security is discussed as a balance between individual rights and public safety, the individual rights in question tend to be presumed to be positive rights (eg., the right of prisoners to go free upon expiry of their sentence) and the role of state power is elided by the discussion of ‘community’ safety. Yet, the nub of the issue, Zedner argues, is the question of ‘[h]ow to enhance security’ against dangers ‘without diminishing security against the state’: how to achieve that, she suggests, ‘is a central problem of our times.’ (2005: 532, emphasis added).
15This analysis of security in the liberal governmentality of contemporary western societies suggests that while mechanisms of security and punishment may overlap in character – such as in the use of penal confinement for public protection – it cannot be assumed that security measures will be regulated by the same constraints as apply in the penal field. Indeed, part of the problem observed by Foucault and Zedner alike concerns the over-reach of mechanisms of security. Zedner (2003a) suggests that the principles of limitation developed in respect of punishment might usefully be extended to the new regimes of security, such as the post-sentence preventive detention laws now found in Australia.
16This makes sense within the broader theoretical frame of liberal governmentality described by Foucault. The freedoms that underwrite liberal government operate, he suggests, through the principles of self interest, and it is the necessary limitations upon such interests that constitute mechanisms of security. Similarly, he claims, in the space of liberal governance the jurisprudence of law will establish interest as the limiting principle of punishment: ‘Punishment will be rooted only in the play of the interests of others, of the family circle, of society, and so on’ (2008 46). Thus the questions asked in liberal jurisprudence will be mimetic of those that actuate government itself: ‘Is it worthwhile punishing? What interest is there in punishing? What form must punishment take for it to be in society’s interests to punish?’ (Ibid). In the present context of mechanisms of security directed at sex offenders, it will be important to understand the jurisprudence of risk that develops around the interplay of governmental and individual interests in the face of danger. To what extent, we may ask, is this new domain of confinement regulated by the principles of limitation that frame punishment – Zedner’s ‘harvesting … analogous principles’ (2003a: 176) – or alternatively set free of constraint, opening up the potential for ‘liberogenic’ effects wherein ‘devices intended to produce freedom … potentially risk producing the exact opposite’ (Foucault 2008 69)?
17The remainder of this paper will consider the way Australian judges have negotiated this difficult nexus, one that lies also at the heart of the country’s sex offender supervision and detention schemes. Our concern here, therefore, will be with the response of Australian judges to the problem of imprisonment when it is framed within the new discourse not of punishment but of security. Their framing of the problem of punishment without a crime, of the demands of community interest and of the dangers posed by sex offenders will come together in what I term here an emerging jurisprudence of risk.
18It will be worthwhile beginning this analysis with a sense of how those seeking the passage of such bills – Australian parliamentarians – have represented the sex offenders with whom they are concerned. All Australian legislation follows with minor variation the model of Queensland’s Dangerous Prisoners (Sexual Offenders) Act 2003. In each state, the legislation has met with little by way of sustained attention within Parliament. In the New South Wales Legislative Council, for example, where after very brief debate the Crimes (Serious Sex Offenders) Bill was passed by a 19-5 vote margin, one member sought to extend the then Premier’s description of criminals as ‘grubs’ by claiming he did ‘not want to refer to these characters as persons, people or human beings’ and that ‘the use of the term “people” is too kind’ (Hansard NSW, 2006: 21812). Against the claim of citizens’ rights to freedom, another member chastised those in the house ‘who appear to have an unnatural and bizarre need to set themselves up as the guardians of sex offenders’ (p. 21805). In the Victorian Parliament the question of rights was situated within a framework of ‘exceptional’ behaviour that would in turn produce exceptional risks and thus exceptional obligations upon the Parliament. ‘Extraordinary crimes’ argued one member ‘must be subject to extraordinary measures’ (Hansard Vic, 2005: 21), while the minister introducing the bill reminded members that the community had ‘a special obligation to our children … and it warrants whatever it is that might be adversely impacted upon in relation to the supposed rights of offenders’ (p. 36, my emphasis).
19Clearly this is not an encouraging start. Yet on the other hand it might be argued that such comments are typical of retrograde back bench members of parliament and should not be seen as reflective of parliamentary attitude more broadly. But that argument is unconvincing on a number of counts. First, not only had the Premier of the state of New South Wales begun the characterisation of criminals as ‘grubs’, but key parliamentary processes concerning the NSW legislation were dispensed with under his leadership. In particular, important questions regarding the legislation’s trampling of liberty rights were raised during select committee stage and the committee expressly asked for parliament to clarify its impacts in this respect. In fact the responsible minister never presented these questions to parliament and the legislation was passed forthwith and absent any reference to the select committee’s concerns. Secondly, the speed with which legislation in each state passed through the house suggests not so much haste as a distinct lack of interest in the detail of measures directed at retarding the ‘supposed’ liberty rights of sex offenders. Finally, all of this is made more problematic by the weight and significance that higher courts in each state has placed upon the ‘intentions’ and ‘wishes’ of parliament. As following sections of this paper will illustrate, Australian courts have gone to great lengths to carve out ponderous and sometimes mysterious formulations that allow them to evade and bypass normal criminal justice standards in order to give effect to governmental wishes that the ordinary citizenship rights of sex offenders should not be restored following completion of a penal sentence.
20Nevertheless, if Lucia Zedner’s (2003a) suggestion that the limiting principles of punishment (eg., of proportionality, parsimony, restraint, or equal impact) might be useful also for delimiting mechanisms of security, then we might expect Australia’s higher courts of appeal to be the place where the first moves towards such a jurisprudence of risk and security would emerge. The case on which I concentrate here is Director of Public Prosecutions (WA) v GTR, a matter heard on appeal in the Western Australia Supreme Court between June and September 2008 that was significant for its breadth of attention to some of they key interpretive questions arising in the WA statute. However, since the drafting of legislative provisions around the country is nearly identical, the Court in GTR also drew upon and discussed a number of other cases of appellate review of sex offender legislation in Western Australia, Queensland, NSW and Victoria, thus providing an Australia-wide relevance. I wish to draw out three points from GTR. These are: (i) the interpretation of the Court as to the evidence necessary in order for a continued supervision or detention order to be made; (ii) the possibility of making no order should the government be unwilling or unable to make good the preventive (treatment/proper supervision) element of the order; and (iii) its decision on the admissibility of evidence of offending as a child. Together, these three factors reveal a court shifting back and forth between criminal and civil standards and precepts in order to establish an entirely new field of legality that makes possible the governmental objectives of quarantine and exclusion.
21The Western Australian legislation sets out in s 40 that proceedings under the Act ‘are to be taken to be criminal proceedings for all purposes’. Further, individuals detained under the Act will be treated as prisoners, being detained ‘in a prison under the Prisons Act 1981’ (s 45). One might expect therefore, given the criminal nature of proceedings and the effective imprisonment under consideration, that a criminal standard of proof would be required in order to secure post-sentence detention or supervision of an offender against predicted future conduct. In fact, in line with other state’s schemes, the WA law establishes a complex set of intersections between standards of proof, estimates of likelihood and vague invocations of seriousness of imagined future conduct. Just how these are worked out, and in whose benefit presumptions and similar legal devices are established, will go a long way in determining the capacity of Australian judges to take seriously the importance of individual freedom for members of ‘suspect’ populations.
22Section 7 of the WA Act establishes that the court must ‘be satisfied that there is an unacceptable risk that, if the person were not subject to a continuing detention order or a supervision order, the person would commit a serious sexual offence’ and that this satisfaction must be ‘to a high degree of probability’. The WA court had previously considered this in Williams , while the ‘high degree of probability’ element had been considered in Victoria in TSL  and in NSW in Tillman  and Cornwall , both of the latter following TSL. The Court in TSL prefaced its conclusions by noting that ‘[a] person subject to an extended supervision order is a prisoner in all but name’ (at 10). Nevertheless, it concluded in respect of the ‘high degree of probability’ question that ‘[t]here is no reason to think that it must be more than 50%.’ (at 11), meaning that while the court must be satisfied that a further serious offence is likely, it should not be required to be more likely than not. The Court in GTR followed this lead, denying that the legislation (and legislative intent) could be construed to mean that, overall, there must be a better than even chance (> 50%) that a serious sexual offence would occur if an order were not made. The language used is that the legislation conveys ‘no requirement’ for such a high standard (at 34). Thus despite the clearly criminal context and outcomes, a civil standard comes to applied as a strategic tool in conditions where no serious hope could be found for success under any measure of criminal (justice) standards. In order to achieve this the courts must take leave of common sense interpretations of language (what part of the term ‘high degree of probability’ does not connote to the ordinary person a probability far greater than 50%?) and establish something else in its place. In doing so it joins an emergent trend in Australia to replace common language, and particularly statistical concepts like probability, with convoluted constructions, such as in the case of Darkan  in relation to the crime of murder. In GTR the Court was certainly cognizant that an abuse was being done to ‘the plain meaning of the language’ (at 34) and so an attempt was made to craft an elaborate structure of logic that would do justice to legislative intent while recognizing the implied demands of language.
23It did so by setting out its interpretation of the cascading decision structure in the WA legislation. The Court in GTR noted (at 34) that for an order to be made there must be (i) satisfaction to a high degree of probability; that (ii) there was an ‘unacceptable risk’; that (iii) a serious sexual offence ‘would’ occur unless a continuing detention or supervision order were made. We are thus left with a proportion (the high degree of probability), of a further proportion (the risk), of an event (the imagined offence), which is itself conditioned by the impact of another event (treatment to be received in detention or on community supervision). Despite the apparently satisfactory elegance of such a set out, any cursory attention to it reveals the whole structure to be a sham. The mathematician will point out that small fractions multiplied (where the satisfaction probability meets the event risk/probability) leave us with something much less. The social scientist will follow Floud and Young (1981) in observing that imagined worst case scenarios (the offence that ‘would’ happen) are a poor basis for prediction or public policy. Once the worst is imagined there tends to arise a fallacious belief that the public should be protected from the risk of that thing occurring. And within all this, others will observe that offenders’ right to liberty appears only as part of a ‘balancing exercise’ (at 27). This, essentially, forms the mechanism by way of which such rights are traded away against the impacts of offenders’ imagined future behaviour, the result going to form the ‘unacceptable risk’ component above.
24What we find in the emerging jurisprudence of risk is an eschewal of the very principles of justice that lie at the heart of traditional criminal judgment and an acceptance of a new hybrid domain of exception, neither fully criminal nor fully civil but construed and constructed so as to give effect to parliamentary intentions for those who may be described, to borrow from Franz Fanon (1967), as the sub-men who are barely human. In this respect it might be well to recall conclusions drawn almost a generation ago, at a time when dangerous offender anxieties were focused on another group, violent offenders, about the aims of public policy. ‘The mere fact of a future outcome’, Floud and Young observed, ‘is not what determines the justice or injustice of preventive measures. The objective in dealing with ‘dangerous’ offenders is to find the just course and not simply the practical way out of the difficulty they present.’ (1981: 60).
25The fundamentally punitive character of continued detention and supervision in Western Australia emerges further when the Court considers the question of whether or not, upon a finding of danger by the process above, a court may decline to make an order. In the Queensland case of Francis  the court there used the phrase ‘executive repudiation of the preventive objects of the Act’ (at 31) to describe government refusal to provide treatment to a willing offender, thus making continued detention ‘truly punitive in character’ (at 31; emphasis added). Leaving aside the very high bar of ‘refusal’ to provide treatment, as opposed, for instance, to provision of insufficient or ineffective treatment or supervision, we face here a very important question. It asks whether a government is bound to effect prevention by every means at its disposal (through treatment, supervision conditions, etc) or whether danger in and of itself may trigger control if a government declines to execute preventive measures, even though it has the power to do so. In GTR the Court found in favour of the latter (at 51), giving enormous license to government to establish minimal, ineffective or even sham measures for the very treatment component that is held to render detention and supervision non-punitive and thus outside or beyond established principles for the limitation of punishment. This would also seem to leave ample scope for poor resourcing of community supervision so as to increase the likelihood that courts will be forced to grant detention as the only reasonable option for securing public safety.
26The possibility of such a circumstance is far from moot. Even as the WA Supreme Court met to consider the appeal in GTR and to regard this issue at the level of principle, its counterpart court in New South Wales was being asked to consider exactly this problem arising in practice. At hand in the appeal case of Brookes  was the question of whether the penal bureaucracy could reasonably withhold resources for community supervision of a sex offender and in so doing render continued detention a fait accompli. The reasoning of the NSW Supreme Court in Brookes is instructive and provides something of a balance to the WA court’s perception that its role was principally to give effect to parliamentary wishes.
27The particular circumstances of Mr Brookes had produced recommendations to the Court that he should receive individual (rather than group) treatment and that it would be satisfactory from a risk mitigation point of view for this to take place in the community. In a counter move, however, the NSW Department of Corrective Services declined to make funds available for such treatment, despite having done so for other individuals in the past. The Department’s Executive Director of Offender Services and Programs explained: ‘I do not consider there to be equivalent special circumstances in Mr Brookes’ case to warrant payment by the Department for private sex offender treatment in the community’ (at 13). Continued detention, it proposed, was therefore the only available option. The NSW Supreme Court’s judgment in Brookes is important for its contemplation of the way risks to community and individual may be variously balanced by supervision or detention options, but perhaps more significantly for its attention to the relationship between parliament, executive and courts.
28Three key points were made. First, the Court in Brookes cited with approval the original judgment of Grove J who observed that a bureaucrat who withheld funding for community treatment effectively made such treatment ‘virtually hypothetical’ (at 14), which clearly could not have been the intention of parliament. Further, he added, ‘I am unaware of the source of a test requiring the demonstration of “special circumstances”’, and still less satisfied that ‘the statutory intent … can be subverted by a bureaucratic decision’ (at 15). The appeal court thus established the principle that the penal bureaucracy should act as a facilitator, not a barrier, to executive intent. Second, on the question of the relationship between such intent and judicial discretion, the Court in Brookes cited an earlier appeal (under the same sex offender legislation) by Winters . There the Court observed: ‘If the Supreme Court was bound to make an order for continued detention, whenever the Executive made no provision for appropriate supervision, it could be said that the Supreme Court was a mere instrument for Government policy’ (at 25). Since that would breach the doctrine of separation of powers, the scope for judicial discretion must lie at some point between blindly giving effect to government policy, on the one hand, and directing government on how to allocate its resources on the other. Finally, the Court in Brookes accepted the formulation of Grove J in the original judgment wherein this impasse was solved by releasing Mr Brookes on extended supervision and directing him to accept treatment only ‘if it is made available’. This construction shifted the onus of action away from the offender (to attend treatment) and on to the penal bureaucracy (to provide it). Only if such treatment were made available would a requirement for Mr Brookes to attend be triggered. This might seem a rather minor inflection, but an important principle is established here, and it is one that reflects Zedner’s notion of harvesting analogous principles from elsewhere in law. Refusing the idea that it was ‘impermissibly’ intervening ‘to dictate the decision’ on allocation of resources, the Court in Brookes observed that once having established that funds ‘could be provided’, it was ‘entitled to expect’ that the responsible bureaucrat would ‘in the reasonable exercise of the discretion given to her, decide to release the funds’. It concluded: ‘In the language of administrative law any other decision would be irrational. The Court must assume that the Executive will not make an irrational decision’. The appeal against Mr Brookes’ release upon extended supervision was thus dismissed and an important protection against overbearing state power established.
29Returning to the case of GTR, however, we find another long held principle of justice extinguished. In this case and despite it not being a matter of appeal, the WA Court felt it important to clarify that the Act’s demand for all material relevant to establishing a risk of further serious sexual offending should over ride provisions of the Young Offenders Act (WA) 1994 that effectively expunge youthful convictions after a period of two years. Here, then, is the question of the admissibility of childhood offending, something long prohibited in criminal law. The Young Offenders Act instructs in s 189(2) that, once two years have passed, ‘the conviction is not to be regarded as a conviction for any purpose’. But it was set off in GTR against the Dangerous Sexual Offenders Act counter claim in s 38(4) that information to establish risk should be supplied ‘despite any other law or duty of confidentiality’. The original judgement in the matter by McKechnie J had in fact ruled in favour of excluding the evidence of youthful conduct. He argued that the prohibition must hold: ‘To hold otherwise is to make effectively meaningless the rehabilitative policy behind the Young Offenders Act s 189 and s 190. To do otherwise would involve a court in a highly artificial process.’ (cited at 92). Consistent with the demeaned status of offenders subject to these proceedings, the Court of Appeal in GTR ruled that the conflict should be settled in favour of the control legislation. In exactly the kind of artificial process imagined by McKechnie J, the Court of Appeal held that ‘this is not a case of the implied repeal of the provisions of the Young Offenders Act by the provisions of the DSO Act.’ Rather, it argued, ‘[t]he provisions sit together in a way which makes the evidence admissible, thereby advancing the purposes of the DSO Act’ (at 114) whilst apparently doing no harm to the Young Offenders Act. Clearly, this presumptive shift was not the first of its type in the judgment as the court made movements back and forward to achieve governmental ends and to head off criminal standards and notions of justice. Finally, even the Court of Appeal itself was forced to admit it could ‘appreciate of course the artificiality of that result’ (at 119).
30This analysis of the jurisprudence of Australian courts in respect of the post-sentence involuntary detention of sex offenders grounds theoretical discussions of liberal governmentality, freedom and security in the data of legal judgement. It provides an empirical test of certain propositions. The central one tested here has been Lucia Zedner’s suggestion that courts, in considering questions of the appropriate justification, allocation and distribution of mechanisms of security, might look to the criminal law for guidance and harvest ‘analogous principles’ (2003a: 176) from the penal domain. This study shows that Australian appellate courts have steadfastly refused that opportunity.
31Indeed, across four state jurisdictions – New South Wales, Queensland, Victoria and Western Australia – the evidence is that when asked to contemplate imprisonment for ‘security purposes’, appellate courts have studiously evaded the principles of justice that shape and limit the use of imprisonment for criminal punishment. Despite evidence that state legislatures gave scant attention to the detail and construction of post-sentence detention schemes, courts have continually fallen back upon arguments about giving expression to ‘parliamentary intent’ as they crafted dubious, unjust and striking illiberal formulations that allow them to evade criminal standards. Despite the fact that sex offenders caught up in these new mechanisms of security are imprisoned in prisons, largely subject to the same penal regime as other prisoners and regularly denied the very treatments (eg, individual therapy) that would allow them adequate preparation for release, Australian courts have eschewed any implied demands to deal with sex offenders on the basis of those standards that would apply in a criminal trial.
32The focus of this paper has been the case of Director of Public Prosecutions (WA) v GTR  wherein a number of these issues were aired. What have not been mentioned thus far are the specific details of the case. It seems appropriate now to mention these circumstances, for they place in context the scope and impact of this new form of preventive control in Australia. The respondent, GTR, was a young Aboriginal man from a remote community. In prefacing his judgement on the original application for supervision, Director of Public Prosecutions v GTR , McKechnie J remarked that of 15 applications thus far made by the Director of Public Prosecutions, seven had concerned Indigenous respondents. Thus, he noted, though Aboriginal people made up just 3% of the WA population, they had thus far been represented in 47% of Dangerous Sex Offender applications. Further, fully 33% of applications had been made in respect of residents of remote communities. The nature of GTR’s offending is also important. Without diminishing the significance of any offending, it must be noted that his comprised just two events, the first of which, committed as a 17 year old, resulted in seven months imprisonment; the second, occurring approximately one year following his release was met with a sentence of three years, four months. Both offences were against adult women. He had undertaken treatment on his most recent sentence and it had been considered successful. McKechnie J, in the original judgement decided that although there was ‘a pattern to his past offending’ there was ‘no evidence of a propensity to so offend’ (at 132), a view supported by the appellate court. The danger with schemes such as these, particularly when constructed and interpreted in the manner described here, is that they open up what Robert Castel (1991: 289) has termed ‘a vast hygienist utopia’ in which the play of community anxieties quickly subsumes any supposed liberty rights of offenders and in which arguably ‘run of the mill’ offenders, such as GTR, may find themselves redefined as a chronic social threat. Thankfully in this case, and despite two applications from the Director of Public Prosecutions, the Western Australia Supreme Court declined to agree.
33Nevertheless, Foucault’s suggestion that the principle of interest that drives liberal governmentality, and which he so neatly described in relation to punishment (2008: 46), operates as a general principle is left somewhat in doubt. This is important, because his elevation of interest was unequivocal: ‘Governmental reason in its modern form’ he proposed ‘is a reason that functions in terms of interest’ (2008: 44). Moreover, the way interest operated in the penal realm was no minor feature of liberal governmentality: rather, he described it as ‘a striking example’ of the interplay of interests (p. 45) underpinning liberalism in toto. We are thus left to question whether the subordination of security to liberty proposed by Foucault is correct or whether, perhaps, as Mark Neocleous (2007) has suggested, the activating principle of liberalism is in fact security’s pre-eminence over freedom.
34Also at issue is whether this new jurisprudence of risk marks out a retreat of law from the field of security. In its most radical formulation, and drawing on the work of Giorgio Agamben (1998, 2005), this argument proposes that sex offenders (the modern homo sacer) come to be placed beyond law entirely. Dale Spencer (2009), for example, returns to the old binary of normal and abnormal, characterising the sex offender, inter alia, as ‘the instantiation of the opposite’ (p. 226). By dint of his un-likeness to us and his instantiated violation of all that we hold sacred, the sex offender becomes subject to the ban, a sovereign declared ‘state of exception where law is suspended and bare life is abandoned by the law’ (Ibid). Somewhat paradoxically, Spencer argues that the use of civil standards and definitions in the case of sex offender preventive detention is equivalent to the ban: ‘the homo sacer qua sex offender is banned from the community. By juridical/institutional decision, the laws protecting his constitutional rights are suspended so that the sovereign can act upon the sex offender’ (p. 229). Unfortunately, this incoherence (shifting legal standards evidencing no law at all) stretches Agamben’s thesis to breaking point. Moreover, the examples given by Agamben himself only go to reinforce this misapprehension of his argument. To illustrate that he is not talking about the periodic ebb and flow of legal standards Agamben gives the U.S. Patriot Act 2001 as a key example. Here, President Bush’s sovereign decree ‘radically erases any legal status of the individual [terror suspect], thus producing a legally un-namable and unclassifiable being’, rendering them ‘objects of a pure defacto rule’ (2005: 3), ‘removed from law and from judicial oversight’ (p.4). The only analogous example Agamben can find – ‘the only thing to which it could possibly be compared’ – is the Jews in Nazi Germany ‘who, along with their citizenship, had lost every legal identity’ (Ibid.).
35The new jurisprudence of risk thus does not mark a retreat of law from the domain of security and is of an entirely different order to that described by Agamben in his notion of the state of exception. What is required here is a theoretical conception of the changed legal rationality: for the disjuncture where rules governing imprisonment as punishment suddenly appear irrelevant to imprisonment used as a tactic of security. One avenue may be to think of this new jurisprudence of risk as marking out a new phase in the constitution of legal subjects. Some time ago I suggested that sex offenders have become bound up in a discourse not of rights (that may be claimed) but of obligations (that are owed), based upon an exclusionary logic inherent within liberalism (Brown, 2005). The origin of the logic, I suggested, can be found in liberalism’s constitution in the political field of colonialism and the tendency therein to constitute colonial subjects not as recipients of rights (as were their metropolitan counterparts) but instead as agents of obligation, the main index of which would be aspects of character and virtue. Yet this thesis, also, is insufficiently specific to do more than account for the general tactics currently applied against sex offenders in the domain of security.
36A framework for understanding the legal refashioning of appropriate norms such as we have observed in relation to post-sentence detention of sex offenders might, however, be found in what Nasser Hussain (2003) has termed the ‘jurisprudence of emergency’. Like Agamben, Hussain begins with Carl Schmitt’s famous definition of the sovereign as ‘he who decides on the exception’. For Hussain this serves as a prelude to examining the place of legal suspension (eg., rule by martial law) or legal reconstitution (eg., adding emergency powers within ordinary law) in the context of danger and emergency. In both cases the doctrine of necessity connects law to danger and provides the grounds for legal improvisation that ultimately ‘threatens something fundamental about law in general’ (p. 106). Nevertheless, Hussain is also at pains to observe that emergency powers should not be seen as opposite to law or to mark an absence of law, ‘for those exceptional powers themselves emerge out of a system of a rule of law’ (p. 135). Law and necessity thus exist in a kind of uncomfortable equilibrium. One the one hand the notion of ‘pressing danger’ which constitutes the category of ‘necessity’ would suggest a temporally short interruption to normal law. Yet any analysis of emergency shows that ‘pressing’ dangers often rapidly bleed into something more akin to a state of permanent emergency. Similarly, although rule of law is designed to regulate and constrain the demands of necessity and particularity, attempts at accommodation within normative legality do ‘not always match the needs of the sovereignty of the state’ (p. 136). In the present context, the efforts of Australian courts to create malleable legal norms, shifting between civil and criminal standards, divorcing language from common usage, constantly referring back to the need to give effect to parliamentary intent, can be understood as a form of suspension of legal normativity in service of necessity. Indeed it is precisely in these notions of pressing danger, of public distress, of risks posed and of governmental obligations to protect, that this new jurisprudence of risk finds its place in marking out the contours of a new security state.
This grew out of presentations made to the Australian and New Zealand Law and Society Association conference in Sydney, Australia, and to the conference Kritisch-Kriminologische Zeitdiagnosen: Pre-Crime und Post-Criminology? in Vienna, Austria. The paper has been improved by comments received at those conferences, by the assistance of Alex Steel and by the responses of two anonymous referees at Champ Penal. Research toward this paper was supported by a Discovery program grant from the Australian Research Council (DP0877331).