University of Birmingham The ‘inside’ and ‘outside’ of prisons : Carceral geography and home visits for prisoners in Finland

. The


Introduction
The term 'carceral geography' (Moran et al. 2012(Moran et al. , 2013Philo 2012) describes a new and vibrant sub-discipline of human geography research into practices of incarceration. Such work is often informed by and in dialogue with the work of Foucault (1979) on the development of the prison, surveillance, and the regulation of space and 'docility' of bodies, and of Agamben (1998Agamben ( , 2005 on the notion of a space of exception, where sovereign power suspends the law, producing a zone of abandonment. The distinctiveness of its focus on the spaces of incarceration is, however, matched only by its tendency to draw primarily on case study examples from high imprisonment societies, and this paper argues that in theorising carceral space, carceral geography needs to be more attentive to the varied contexts of incarceration. The paper first discusses the tendency within carceral geography to focus on case studies from high imprisonment societies, and suggests, by discussing the specific context of Finland, that the interdisciplinary dialogue emerging between carceral geography and criminology could be expanded to consider a more active engagement with comparative criminology. Secondly, it critiques Goffman's (1961) 'total institution' thesis by considering some aspects of the furlough, or home visiting scheme for prisoners in Finland, and by determining the extent to which the furlough is an example of the concept of 'heterotopia' (Foucault 1998) deployed by Baer and Ravneberg (2008) to understand the nature of the 'inside' and 'outside' of prisons.

Carceral geography
In the United States (US), Wacquant (2011: 3) has described 'a brutal swing from the social to the penal management of poverty'. The core of Wac-The 'inside' and 'outside' of prisons: Carceral geography and … quant's thesis is the penalization of poverty seen in the US in recent decades. This tendency arguably extends into Western Europe and elsewhere through the exporting of US penal ideas and management systems (Downes 2007: 118, in Gottschalk 2009, comprising a 'punitive revamping' of public policy by tackling urban marginality through punitive containment. Hyperincarceration, having in the United States thrown its 'carceral mesh' (Wacquant 2011: 13) around the hyperghetto, is argued to have established a 'single carceral continuum' between the ghetto and the prison system in a 'self-perpetuating cycle of social and legal marginality with devastating personal and social consequences' (Wacquant 2000: 384). Imprisonment rates in the United States are the highest in the world, but as Aebi andKuhn (2000: 66, cited in von Hofer 2003: 23) point out, such rates 'are to a great degree a function of criminal justice and social policies that either encourage or discourage the use of incarceration' rather than a function of the number of crimes which are committed.
Carceral geography increasingly spans and synthesizes three main areas of interest; the nature of carceral spaces and experiences within them; spatial or distributional geographies of carceral systems; and the relationship between a notion of the carceral and an increasingly punitive state. Drawing upon case studies from South Africa and the USA respectively, Dirsuweit (1999), and Sibley and van Hoven contest the Foucauldian regulation of prison space and the docility of bodies, describing carceral 'spaces… produced and reproduced on a daily basis' (van Hoven & Sibley 2008: 1016, and the agency of inmates making 'their own spaces, material and imagined' (Sibley & van Hoven 2009: 205). Spatial geographies of incarceration have again focussed on the US, with researchers considering the relationship between places of incarceration and the communities which host or surround them (Che 2005;Glasmeier & Farrigan 2007;Bonds 2009) extending critiques of the 'total institution' (Goffman 1961), and suggesting that the 'carceral' is something more than merely the spaces in which individuals are confined. Such a notion of the 'carceral' as a social construction relevant both within and outside physical spaces of incarceration, informs, for example, the work of Allspach (2010) on 'transcarceral' spaces in Canada, and the theorisation of the 'carceral' as embodied through the corporeal inscription of released inmates (Moran 2012a(Moran , 2013 in the Russian Fed-eration. Again in the United States, Peck (2003) and Peck and Theodore (2009) discuss the relationship between prisons and the metropolis in the context of hyperincarceration. Much of this work, then, originates in or pertains to, highly incarcerative settings; the USA and Russian Federation have some of the world's highest incarceration rates, of 730 and 529/100,000 respectively in 2010 (World Prison Brief 2012). Baer and Ravneberg's work (2008) is an interesting example in this context, since it compares a increasingly incarcerative setting, that of the English prison system, with the very low imprisonment society of Norway. Baer and Ravneberg (2008) problematise the conceptualisation of a binary distinction between the 'inside' and the 'outside' of prisons, instead positing prisons as 'heterotopic spaces outside of and different from other spaces, but still inside the general social order' (Baer & Ravneberg 2008: 214). Although the authors note that their two research contexts are different, they stop short of elaborating very far on the marked differences between them, or suggesting that these differences may have implications for their theorisation of 'inside' and 'outside'. This paper therefore pursues two issues raised by this work; first, the issue of the carceral context, which they obliquely raise but do not fully resolve, and second Goffman's conceptualisation of the 'total institution' (1961) which they directly critique via the notion of 'heterotopia'. The first of these is discussed next, in relation to the carceral context of Finland.

Finland's penal culture
Finland is an interesting case study for a number of reasons. Its relatively small prison population and diligent record-keeping enables fine grained statistical research to be carried out, but more importantly, it differs very significantly from the research contexts which have thus far ignited geographers' interest in incarceration -the United States and Western Europe. Whereas the United States is highly incarcerative (or perhaps hypercarcerative), other countries are by contrast decarcerative, actively deploying different techniques to decrease their prison populations. Finland is one of these countries.
As Lappi-Seppälä (2000) has argued, the dramatic decrease in Finland's prison population, from the early 1950s when its imprisonment rate Dominique Moran and Anssi Keinänen of 200/100,000 population was four times higher than other Nordic countries, to the situation sixty years later when its imprisonment rate is now the lowest in Scandinavia at 52/100,000, was brought about by 'a conscious, long-term and systematic criminal policy ' (ibid. 37−38). This policy and the philosophies which underpin it are fundamentally different from those which inform criminal justice policy in, for example, the US. The 1960s and 1970s saw in Finland the re-evaluation of the theoretical aim and the justification of punishment, with a 'turn' towards 'general prevention' rather than direct 'deterrence', through a 'humane neoclassical crime policy'. General prevention essentially means the 'moral-creating and value-shaping effect' of punishment (von Hofer 2003: 28), which differs from deterrence in that individuals are not perceived to obey the law because of the direct fear of punishment, but rather because the disapproval expressed in punishment is perceived to influence the values and moral views of individuals through a process of internalisation of the norms of criminal law. As von Hofer (ibid.) puts it, 'people refrain from illegal behaviour not because such behaviour would be followed by unpleasant punishment, but because the behaviour itself is regarded as morally blameworthy'. This 'turn' was enabled and enacted by an exceptionally expertoriented reform team with close professional and personal contacts between politicians, state officials and academic researchers (Lappi-Seppälä 2002) in the absence of political opposition to reform; by the 'attitudinal readiness' of the Finnish judiciary to adopt a new sentencing policy (von Hofer 2003: 31); and by the 'fairly sober and reasonable attitude towards issues of criminal policy' (Lappi-Seppälä 2000: 37−38) adopted by the Finnish media. In a broader context, for much of this period of decarceration, Finland was not a destination for significant immigrant movement, and was additionally insulated from international trade in narcotics, two major influences on prison populations (Christie 2000).
What this essentially means is that Finland has for decades been decoupled from the US (and increasingly the European) tendency to politicize criminal justice policy to the extent that criminal justice becomes a political tool rather than a balanced assessment of criminal justice interventions.
As Lappi-Seppälä (2002: 33) observes, in these contexts 'the higher the level of political authority, the more simplistic the approaches advocated. The results can be seen in slogans that are com-pressed into two or three words, including "prison works", "war on drugs" and "zero tolerance"' which in turn leads politicians to 'pander to punitive (or presumably punitive) public opinion with harsh tough-on-crime campaigns'. Although Finland has perhaps reached its lowest possible imprisonment rate, with the result that rates will almost certainly rise, and the Finnish media is becoming more active in debating criminal justice policy, the 'humane neoclassical crime policy' continues to prevail; prison is not considered to "work" and the solutions to social problems are not 'sought where they cannot be found -the penal system' (ibid. 33). In the Finnish system, then, it may be argued that prison policy is informed more by an understanding of the likely success of specific interventions for the stated aims of incarceration, than by a political imperative to respond to public opinion, or as Gilmore (2002: 16) has argued, to use the prison system as 'a project of state-building'.
The philosophy towards the conditions of imprisonment in Finland is displayed in The Sentences Enforcement Act 1 which states that Punishment is a mere loss of liberty: The enforcement of sentence must be organised so that the sentence is only loss of liberty. Other restrictions can be used to the extent that the security of custody and the prison order require.
Prevention of harm, promoting of placement into society: Punishment shall be enforced so that it does not unnecessarily impede but, if possible, promotes a prisoner's placement in society. Harms caused by imprisonment must be prevented, if possible.
Normality: The circumstances in a penal institution must be organised so that they correspond to those prevailing in the rest of society.
Justness, respect for human dignity, prohibition of discrimination: Prisoners must be treated justly and respecting their human dignity.
In essence, this means that the 'less eligibility' principle that informs much prison policy in the US and Western Europe, that prisoners should 'suffer' in prison, not only through the loss of freedom but also by virtue of prison conditions, which should be of a worse standard than those available to the poorest free workers, is absent in Finland (Pratt & Eriksson 2011). Instead, prison conditions are to correspond as closely as possible to living conditions in society (Ministry of Justice 1975), with the intention that penalties for offences are The 'inside ' and 'outside' of prisons: Carceral geography and … implemented in such a way that they do not unduly interfere with prisoners' participation in society, but as far as possible, promote it.
Pursuing the second issue raised by Baer and Ravneberg's (2008) work, we next consider critiques of Goffman's conceptualisation of the 'total institution' (1961) which they directly critique via the notion of 'heterotopia'. Goffman's (1961) theorisation of the 'total institution' has been an influential piece of work, which since its publication has seen the concept of a 'total institution', defined as …a place of residence and work where a large number of like-situated individuals, cut off from the wider society for an appreciable length of time, together lead an enclosed, formally administered round of life (Goffman 1961: 11).

The 'total institution'
The concept has been applied to an extraordinarily diverse range of circumstances and contexts, such as homes for the elderly (e.g. Mali 2008), psychiatric units (Skorpen et al. 2008), the home (Noga 1991), the mass media (Altheide 1991), the military and the police (Rosenbloom 2011), and sport (Cavalier 2011).
While the concept has thereby been shown to have considerable utility, its appropriateness as a means of understanding the types of institution in relation to which it was initially developed has been called into question. For the purposes of this paper, it is the applicability of the 'total institution' thesis to the institution of the prison which is most salient, and in this context, there are a number of commentaries which point out some disjunctures between the theory and the actuality of imprisonment. Perhaps the most thorough treatment was that of Farrington (1992: 6), who problematised Goffman's theorisation by arguing compellingly in relation to the US prison system that the 'total institution' thesis is 'in fact, fairly inaccurate as a portrayal of the structure and functioning of the... correctional institution' in that the modern prison 'is not as completely or effectively "cut off from wider society" as Goffman's description might lead us to believe'. Although Farrington (1992) addressed his comments explicitly to this geographical context, and perhaps interpreted Goffman rather too literally, many of his observations are generically applica-ble to prison systems more generally, and they bear some exploration here.
The core of Farrington's (1992) commentary is that prison institutions have a relatively stable and ongoing network of transactions, exchanges and relationships which connect and bind them to their immediate host community and to society more generally (ibid. 7). Although at the time of writing Farrington (1992) observed that relatively little research had explored these connections, such as the relationships between prisons and their host communities, the process of prison siting, and the relationships between criminal offenders and the society from which they have come, in the intervening years these topics have come more clearly into view in academic scholarship, and particularly in the recent development of carceral geography which has considered precisely these issues. The relationship between prisons and wider society has been a particular focus of study (e.g. Peck 2003;Peck & Theodore 2009). Rather than rejecting Goffman's thesis, Farrington (1992: 7) essentially extends and develops the notion of the prison as a 'total institution' by proposing a theoretical conception of 'a "not-so-total" institution, enclosed within an identifiable-yetpermeable membrane of structures, mechanisms and policies, all of which maintain, at most, a selective and imperfect degree of separation between what exists inside of and what lies beyond prison walls.' Subsequent studies support this notion, with, for example, Hartman (2000) discussing the restriction of prisoner access to the internet in the language of 'walls and firewalls'.
Although Farrington (1992) identifies 'points of interpenetration' through which the prison and wider society intrude into and intersect with one another, his critique stops short of that of Baer and Ravneberg (2008) who problematise the conceptualisation of a binary distinction between 'inside' and 'outside'. Instead they posit the concept of heterotopia, viewing prisons as 'heterotopic spaces outside of and different from other spaces, but still inside the general social order' (Baer & Ravneberg 2008: 214), which they argue renders problematic the separation of inside from outside. They build on Foucault's work, in which he characterised heterotopias as 'real places, actual places, places that are designed into the very institution of society' (1998: 178) but which can seem totally unrelated to one another despite existing side by side. In so doing, they rely particularly on Genocchio's (1995) observation that heterotopias Dominique Moran and Anssi Keinänen are 'outside' of and fundamentally different to all other spaces, but also relate to and exist within general social space that distinguishes their meaning as difference. Baer and Ravneberg (2008) argue that the concept of heterotopia allows for a fuller understanding of the spatial complexities of the prison environment than the total institution thesis which distinguishes between inside and outside with very little room for blurring of this boundary. In their comparison of English and Norwegian prisons, they found what they described as 'incompatible juxtapositions' (ibid. 212), in which there were 'multiple, simultaneous distinctions and indistinctions' between the inside and outside of prisons, rather than a set of neat binaries (i.e. inside/outside). They described that they 'sensed a lack of delineation between inside and outside at the same time that there was sharp distinction within prison' (ibid. 213), and that prison seemed 'to be a compressed mélange of seemingly incompatible juxtapositions' (ibid.) with 'tension and fusion between inside and outside' (ibid. 214) These 'incompatible juxtapositions' derive from Baer and Ravneberg's personal impressions of entering and leaving English and Norwegian prisons, or in Farrington's (1992) terms, their individual 'interpenetration' of the penitentiary wall from a position on the 'outside'. Other work within carceral geography (Moran 2012b) which challenges the total institution thesis has also considered the crossing of this assumed boundary from outside to inside, with prison visitors entering visiting spaces 'inside' the prison walls, and has considered the sense in which prisoners coming to meet them experience something of the 'outside' via this contact. In this paper, however, we explore the opposite direction of movement; prisoners from 'inside' going 'outside' the prison, through the practice of home visits or furloughs, and consider the extent to which this practice destabilises the inside/outside binary through the notion of heterotopia.

Theorising prison visiting and furlough schemes
Research into furloughs is part of a wider body of work, largely within criminology, but also prison sociology and more recently carceral geography, which explores prison visitation as a space of interpenetration between 'outside' and 'inside', and which considers the significance of visitation for wider penological concerns such as recidivism, or reoffending after release, and the 'collateral' effect of imprisonment for prisoners' friends and family.
Research into prison visiting per se has tended to orient itself around two foci -the effects of visitation on prisoners and their subsequent behaviour, especially after release, and the collateral effects of imprisonment on inmates' family members, mitigated to some extent by visiting as means of maintaining contact.
Considering first the effects of visitation on prisoners, criminologists studying recidivism find that prison visitation is a significant factor in improving post-release outcomes. Holt and Miller (1972) showed that parole outcomes were much more positive for visited prisoners, and lower recidivism rates have since been demonstrated across study populations and time periods. However, although the effect is widely observed, the causality is poorly understood. It is presumed that the maintenance of personal relationships and the feeling of 'connectedness' to home and community which may arise through visitation smooth reintegration after release, but this process has never been fully explored. The recent work of social theorists and geographers such as Gilmore (2007), Peck and Theodore (2009) and Wacquant (2010aWacquant ( , 2010bWacquant ( , 2010c, in the United States, in the context of calls for greater attention to the causes of and solutions to hyperincarceration (Wacquant 2010b: 74) and the carceral 'churn' (Peck & Theodore 2009: 251), has been paralleled within criminology by reconsideration of the positive effect of visitation. Empirical studies by Bales and Mears (2008), Derkzen et al. (2009), Berg andHuebner (2010) and Mears et al. (2011) still observe a positive effect, but emphasise its complexity. They draw attention to the effects of visitation of different types and in varying amounts, and on various types of recidivism (Mears et al. 2011), and stress the need for further empirical work to develop a more nuanced understanding of visitation. Mears (2008, also Mears et al. 2011) suggest that future studies should use different, more refined measures to engage with effects of specific types of visit (such as by spouses, children, criminal associates), or nature of visits (consistent or inconsistent during a sentence, characterised by calm and supportive conversations, or by argument and recrimination) in order to better understand the dynamics of visitation.
Before this recent rekindling of interest in visitation, the conclusion that prison inmates visited during imprisonment 'do better' on release re-mained relatively unchallenged for almost forty years. Coupled with an increasing awareness of and concern for prisoner rights and welfare, this relationship formed the basis for policy recommendations encouraging increased numbers and frequency of visits, with greater capacity, longer hours, more programmes to encourage visitors, and to mitigate against the factors discouraging visiting, for example the cost of travel (Schafer 1994). Given this consensus of opinion, academic attention veered away from prisoners' responses to visitation, towards the second focus, on the wider effects of incarceration. The work of Morris (1965) was precursor to an explosion of interest, particularly in the US, broadly in parallel with the growth of mass imprisonment (Garland 2001), in 'secondary' or 'collateral' effects of imprisonment. These range from those impacting directly on prisoners' families and children (e.g. Casey-Acevedo & Bakken 2008; Comfort 2002Comfort , 2003Comfort , 2007Comfort , 2008Codd 2007Codd , 2008Hong-Chui 2010;Wildeman & Western 2010;Comfort et al. 2011;Geller et al. 2011;Kruttschnitt 2011;Shedd 2011), to community effects on labour market participation, civic engagement, and community health (Garland 2001;Lynch & Sabol 2003). Rose and Clear (1998, and Clear 2007 have written about the stigmatisation of communities with high levels of both imprisonment and subsequent re-entry of released prisoners, with parents raising their children in areas with little social control, ambivalent attitudes to law enforcement and public authority, and in which incarceration becomes a way of life (Breen 2010). Wacquant (2010a: 611) argues that prisoners in the US do not, in fact, 're-enter' society after release, instead circulating between the two poles of 'a continuum of forced confinement formed by the prison and... the metropolis'.
Relatively overlooked in this literature on prison visiting is the 'home visit' or 'furlough', in which rather than have family and friends enter the prison complex to visit them, prisoners are granted permission to leave the prison and visit relatives and friends outside, for a specified and restricted length of time. Scholarship of furloughs is relatively limited, and as Baumer et al. (2009) have noted, recent research has tended to focus on primarily descriptive studies which identify categories of prisoner least likely to be granted a furlough, or those who are most likely to return to the prison late or not at all (e.g. Cheliotis 2005;Cid 2005). Some earlier work, followed up by Baumer et al. (2009), considered the functionality of furloughs in easing re-entry after a period of incarceration (Jeffrey & Woolpert 1974), finding that prisoners granted release from prison for vocational or family-related purposes were significantly less likely to be reimprisoned up to four years after the end of their sentence. Drawing further upon the example of furloughs in Ireland, O'Donnell and Jewkes (2011: 75) note that decisions about permitting prisoners home leave in the UK and Ireland 'provide insights into divergent penal policies and contrasting socio-cultural attitudes towards prisoners and imprisonment', and they raise questions about the interplay between penal policy in assessing prisoners' eligibility for furlough, and public opinion and justice policy as political instrument, particularly in the UK where tabloid press coverage of 'temporary release' tends to promote a sense of prisoners as 'undeserving' and 'pampered' (ibid. 89). Although these studies of furloughs provide fascinating insights into the functioning of systems of home leave, rarely are theorisations of furloughs offered, in relation to the significance of the movement of prisoners out of prison, rather than the movement of visitors in.
This paper therefore presents empirical material generated in relation to furloughs in Finland, considering the extent to which this practice challenges the 'total institution' thesis, and the usefulness of the notion of heterotopia in understanding it. These data are analysed and interpreted in relation to two queries; firstly, the extent to which the practice of furlough in Finland constitutes an 'interpenetration' of the boundary of the 'total institution', and secondly, the extent to which the notion of heterotopia assists in understanding the granting and breaching of furlough. Using this data we ask whether furloughs could be interpreted as an example of a heterotopic space, simultaneously 'inside' and 'outside' the prison, just as the prison itself is simultaneously 'inside' and 'outside' the general social order.

Methodology
Data presented in this paper were generated by analysis of two datasets collected for the Finnish Criminal Sanctions Agency, which operates under the Ministry of Justice's Criminal Policy Department, and oversees the Probation Service and the Prison Service. Data pertain to the granting and breaching of furlough, or home leave, for prisoners in Finland.
The first dataset presented here (Table 1) comprises a 100% sample of the applications for furlough in 2008, analysed to determine the factors which contributed to successful applications. Data were analysed by logistic regression to produce correlation coefficients with p-values to indicate statistical significance, and odds ratios which can be interpreted as marginal effects. Emboldened coefficients indicate a statistically significant result, at a 5% level (p<0.05), or a moderately statistically significant result (where the p-value is in the range of 0.05-0.1). The second dataset (Table 2) pertains to furloughs undertaken between October 2006 and January 2009, in terms of the types of breaches of furlough conditions recorded, and the contributory factors identified for these breaches. Breach of furlough is defined as late return to prison, or positive test for a banned substance on return. These data were again analysed to produce correlation coefficients, p-values and marginal impacts or relative risks (RR); with bold text again highlighting statistically significant results. In this table, the relative risks represent the factor by which risk of breach of furlough is higher than the average or comparator. In addition to the statistical data, structured interviews were carried out with prison personnel involved in the furlough system, to illuminate the findings of the statistical research.

The granting and breaching of furloughs in Finland
If a prisoner furloughed in Finland complies with leave conditions, then the time spent on leave is considered to count towards the sentence served. In granting furlough, then, the Finnish prison authorities are effectively selecting prisoners to serve periods of their sentences outside of the prison. Prisoners residing in Finland's 26 prisons (12 of which are open prisons) are eligible to apply for furlough once two thirds of their sentence has been served, or after half has been served if this is specified in an individual prisoner's sentence plan (i.e. the intentions for the prisoner's rehabilitation), or if home leave is required for a particularly urgent personal reason. Furlough can be granted for spells of two hours up to three to four days in every two month period, and the travel costs incurred during furlough can be reimbursed from State funds in certain cases. The movement of prisoners outside of the prison to undertake home visits is a systemic and relatively uncontroversial aspect of imprisonment in Finland, and a relative commonplace for many Finnish prisoners; in 2008 11,312 furloughs were undertaken, an average of three per prisoner. In the context of the 'total institution', it would seem that this prison boundary is particularly porous, that most Finnish prisoners are not as 'cut off' from the wider society as Goffman's (1961) thesis would seem to suggest, and that accordingly, the inside/outside binary might be a problematic notion in this context. However, even if this challenge to the 'total institution' thesis is fairly clearcut, understanding the significance of the granting and breaching of furlough is more complex.
Finland's total prison population in 2008 was 3,526 2 , serving sentences on average of nine months duration. From this prison population there were 15,257 applications for furlough. Of these, 3,897 were refused, and 11,336 granted, giving a 'success rate' of around 74%, with the possibility both of multiple applications from, and multiple furloughs taken by, individual prisoners. As Keinänen et al. (2010) have noted in a report for the Criminal Sanctions Agency, the likelihood of a successful application for furlough depends on a number of different factors, a selection of which are shown in Table 1, and these shed some light on the functionality of the furlough in the eyes of those selecting candidates. The 'marginal effect' describes by how many percentage points the predicted probability would change if the independent variable changed. In other words, considering, for example, the probability of being granted furlough by age group, regression analysis shows that for prisoners aged 50 and over, in comparison with the 17−21 age group, the probability of being granted furlough was almost eleven percentage points higher. Prisoners who had served nine or more previous sentences were three percentage points less likely to be granted furlough than those with no previous sentences, women were five percentage points more likely to be granted furlough than men, and so on.
Although furlough is relatively common, access to it is strictly controlled, through an application process. The space of furlough could be conceptualised as heterotopic; a simultaneous distinction and indistinction of inside and outside; a space outside the prison which offers freedom from carceral control, but to which access is strictly controlled by the penal authorities.
The granting of furlough is case-specific, with individual applications considered in relation to prisoners' sentence plans. Factors taken into account include the nature of the offence committed, and the risk of breach of furlough conditions, a judgement based on the number of previous successful or breached furloughs, the perceived risk of recidivism, any records of attempted escapes or unauthorised leave, and any record of participation in (either positive or negative) activities outside prison. According to this data, when granting access to the 'outside' of the prison, the penal authorities rewarded prisoners' 'docility', in terms of their compliance with the carceral regime during incarceration. In deciding which prisoners would be allowed to traverse the boundary of the prison 'proper', they tended to reward good behaviour in prison, with inmates who had committed five or more disciplinary offences during incarceration being nearly 15 percentage points less likely to be granted furlough, and those having committed only one offence still at a disadvantage of nine percentage points. The data presented in Table 1 also reveal a tendency for penal authorities to view older prisoners, first-time inmates, women, Finnish nationals, and married prisoners more favourably when granting furlough. Although these categories do not map directly onto prisoner behaviour, they suggest that the penal authorities perceived a lower degree of risk of breach of furlough conditions on the part of older prisoners, generally considered to pose less of a security risk (Snyder et al. 2009), alongside 'lower-risk' offenders such as those serving a sentence for the first time, home nationals rather than foreign immigrants, and those who were married and might therefore be perceived to 'comply' with societal 'norms'. This suggests that in the selection process, penal authorities attempt, in the context of the furlough, to blur the distinction between the inside and the outside of the prison, by selecting candidates who they perceive to be more likely to sustain the good behaviour demonstrated inside the prison, on the outside. In so doing, one could argue that they attempt to hold the outside and inside in tension and fusion with one another, granting access to the outside on the basis of good behaviour, and encouraging the translocation of 'docility' demonstrated inside the prison, into the world outside. Or in other words, that they attempt to create, through the case-by-case selection of prisoners for furlough, a space that is simultaneously 'outside' of the prison, but which when properly observed, exists both 'within' the prison (since the prison sentence continues to be served), and within the general social order of the 'outside' (after Genocchio 1995, in Baer & Ravneberg 2008). The extremely low rate of breach of furlough conditions in Finland (less than five percent across the entire prison population) suggests that this selection procedure is relatively successful, allowing only those prisoners at low risk of breach to engage in home leave. Further analysis of the dataset attempted to identify the relative risk of breaches of furlough conditions according to different prisoner characteristics, to see if selection criteria and risk of violation corresponded with one another. Some of these findings bear out the judgements of the panels granting furlough, about the likelihood of breach, but some call them into question. Older prisoners were more likely to be granted furlough, and Table 2 shows that the likelihood of breaching furlough for the over-50s is 0.35 times (or 65% less than) that of 17−21 year olds. Likewise, judgements based on disciplinary record in prison are supported by the finding that prisoners who have committed two to four disciplinary offences during imprisonment are 1.5 times more likely to breach furlough than those with a clean record. Unmarried prisoners are 1.6 times more likely to violate the rules than married inmates, and recidivists are both less likely to be granted furlough, and more likely to violate it.
However, other findings suggest that the selection panels' judgements on risk of furlough breach are not supported by the evidence of the likelihood of violation. Women, for example, are five percentage points more likely than men to be granted furlough, but 1.78 times as likely to breach the conditions of the home visit. Although violent criminals are 4.75 percentage points less likely to be granted furlough, they are statistically no more likely to breach its conditions. Foreign citizens are almost ten percentage points less likely to be granted leave, but no more likely to breach it than home nationals. Prisoner behaviour during furlough seems, therefore, to be rather unpredictable, and the selection process may not entirely account for the low overall rate of furlough violation.
Another explanatory factor could perhaps be that the systems put in place to monitor furloughs serve to successfully deter prisoners from committing violations, and to maintain the heterotopic nature of these spaces. Specific techniques of discipline are deployed, which recall or replicate some of the systems of control used within the prison. These techniques act directly on the body

Un/Supervised
Furlough with a guard -2.675 0.000 -3.21 0.07 Source: Keinänen et al. 2010: 46−54, 74. of the prisoner, requiring physical examination to check for substance abuse, and demanding certain actions on the part of the furloughed inmate in order to comply with regulations. In some cases they act to punctuate the prisoner's passage between inside and outside and back again, highlighting the sharp distinction with the prison. For example, breathalyser and urine tests are administered before prisoners leave the institution, and again when they return, marking their corporeal exit Dominique Moran and Anssi Keinänen from and entry to the prison. However, whilst outside the prison, mobile phones are monitored, with calls from the prison expected to be answered, and requirements to call the prison at set times of day. In some circumstances, furloughed prisoners are also expected to report to local police stations to undergo breath tests. Prisoners whose offences relate to child protection are escorted on furlough by prison guards, who themselves embody the disciplinary regime of the prison, beyond the prison walls. In this way the techniques of discipline also serve to blur or obscure the delineation between inside and outside. The furlough thus becomes a heterotopic, quasicarceral space outside of the prison; access to the 'outside' is not only strictly controlled by the penal authorities' selection procedures, but the 'outside' is actively surveilled and prisoners on furlough are constantly reminded of their incarcerated status.
Interviews with prison personnel question the effectiveness of these techniques of discipline for maintaining the juxtaposition of inside and outside during furlough. Some did feel that the technical measures had some kind of deterrent effect, but others took the view that these measures only made proving a violation easier, and did not of themselves prevent it from happening. We may argue, therefore, that the low rate of breach of furlough is due to the internalisation of the carceral regime, as inmates operate the self-discipline or self-surveillance described by Foucault (1979) as the mechanism through which disciplinary power or biopower produces 'docile' bodies within the carceral institution. This internalisation is represented by good behaviour on the part of prisoners, and operates in two ways in relation to furlough. First, this internalisation is rewarded through the selection process; the penal authorities select 'docile' prisoners, viewed as pliant and most likely to maintain self-discipline, to ensure that furlough is properly observed as a space that is simultaneously 'outside' of the prison, but also 'within' the prison for the purposes of discipline and the contiguous serving of sentences. These individuals are least likely to take advantage of the indistinction of furlough, by acting as if they were 'free'; rather retaining a sense of their incarceration and thereby complying with regulations. Second, even in spaces 'outside' the prison walls, these furloughed inmates continue to operate the self-regulation learned 'inside', complying with furlough conditions and returning to prison sober and on time more than ninety-five percent of the time.

Conclusion
This paper opened with a critique of the relatively narrow empirical focus of carceral geography, drawing attention to its concentration thus far on highly incarcerative contexts, such as the US, UK and Russia, and suggesting that much could be gained from considering different types of penal regime, such as the low imprisonment context of Finland. It then moved to a summary of critical commentary on Goffman's (1961) total institution thesis, drawing particularly on the work of Baer and Ravneberg (2008) who questioned the inside/ outside binary of the total institution through a consideration of the notion of heterotopia to describe the multiple simultaneous distinctions and indistinctions, and incompatible juxtapositions between inside and outside. By positing the practice of furloughs or home visits for prisoners in Finland as an example of the porosity of the prison boundary, the paper argued that the granting and breaching of furloughs could be interpreted as an example of such a heterotopic space, simultaneously within and outside the prison, just as the prison itself is simultaneously within and outside the general social order.
However, we also argue that the context of incarceration is key. Although Baer and Ravneberg (2008) provide a thoughtful and informative theorisation of the inside and outside of prisons which critiques the total institution, and which can be usefully applied to Finnish furloughs, in their own work their focus is turned away from their comparative contexts of England and Norway. We argue here that the specific carceral context of Finland matters very much. As discussed earlier, the Finnish prison system is not intentionally punitive -it aims to treat individuals on a case-by-case basis and to develop a 'moral-creating and valueshaping effect' (von Hofer 2003: 28), based not on punishment or fear of punishment, but by influencing individuals' values and moral views. Importantly, it aims to minimise the differences between living standards in prison and on the outside, by ensuring that loss of liberty is the only 'punishment' suffered. The philosophy that underlies this decarcerative penal system, in which more than a quarter of prisoners are held in open prisons, is that even though imprisonment is intended to include the 'punishment' of loss of liberty, the harm caused by this loss is also to be mitigated as far as possible, by providing support for prisoner adaptation, in the form of educational and work programmes, probationary release prior to the end of sentences, parole, and home leave, or furlough. Furloughs are provided for in Chapter 14 of the Finnish Penal Code, and are specifically intended to support the maintenance of prisoners' social contacts, and contact with wider society, and thereby to reduce the harm caused by loss of liberty. In this context, where conditions 'inside' prison are intended to mirror those 'outside' as closely as possible, where considerable efforts are made to maintain prisoners' social contacts, and where every year the prison boundary is traversed thousands of times by prisoners on home leave, we might expect the 'total institution' as commonly characterised, to fit particularly poorly. In Finland, furloughs can be described as heterotopic spaces, simultaneously within and outside the prison, perhaps because in Finland prison itself occupies a space which whilst necessarily 'outside' the general social order, is also intentionally designed to be very firmly 'within' it, in relation to the stated intention of the Finnish system that imprisonment, should not unduly interfere with prisoners' participation in society, but should as far as possible promote it.
The developing sub-discipline of carceral geography is in active dialogue with more longstanding disciplines engaged with the study of incarceration, and we argue here that comparative criminal justice, the study of 'what people and institutions in different places do -and should do -about crime problems' (Nelken 2010: 1) could be particularly informative. Human geography continually holds in tension the generic and the specific; an imperative to theorise with an awareness of local contingency. Comparative criminal justice similarly tries to consider carceral practices in one place in the light of those in another in order to move towards a more holistic understanding of how crime and its control are connected, and to assist in formulating and testing explanatory hypotheses, and generating transcultural knowledge. In expanding its reach beyond hypercarcerative settings, carceral geography could benefit from an interdisciplinary dialogue with comparative criminology. The current focus within carceral geography on highly incarcerative contexts such as the US and Russian Federation, while offering useful insights into the functioning of these carceral systems and experiences within them, is at risk of encouraging theorisation about carceral systems in general, based on these highly specific settings, and to the exclusion of alternative perspectives and contexts. Whilst considering and comparing different carceral contexts is a challenging prospect, particularly given the frequently cited difficulties of access to prisons and criminal justice systems, we argue here that carceral geography need not start ab initio; prison scholarship recognises that punishment and crime have little to do with one another, and has developed persuasive accounts not only of the workings of criminal justice systems in different places, but of the making of sense of these systems, and the role of researchers in constructing discourses about them. With this in mind, we would encourage geographers working in incarceration to pursue actively interdisciplinary engagements with carceral space, and in particular to draw upon these comparative criminological literatures.