The struggle is real: Theorising community justice restructuring agonistically

This article is a novel use of the ‘agonistic framework’ – a theory of penal change developed in the US, which emphasises the role of hidden conflict – to analyse recent organisational reforms to probation in Scotland. It begins by drawing on recent empirical data to analyse the role of conflict between centralising and localist interests in driving these reforms. This is contrasted with a Scottish policy consensus over decarceration through diversion to community penalties, which despite broad support has been unsuccessful. To explain this contradictory situation, the article builds on recent agonistic literature on the exclusion of some conflicts from penal fields, adding new insights about the circumscription of smaller penal fields. It argues that together these factors serve to ‘crowd out’ debates necessary for substantive change. This new development of the agonistic framework helps explain Scotland’s lack of progress towards decarceration, with policy relevance for other smaller jurisdictions.


Introduction
This article develops contemporary theoretical work in the sociology of punishment by using the insights of the 'agonistic framework' (Goodman et al., 2015(Goodman et al., , 2017 to examine a recent set of probation governance reforms in Scotland. Under the 2016 Community Justice (Scotland) Act, the eight regional Community Justice Authorities (CJAs) were abolished and replaced by a two-tier system of local partnerships alongside a national leadership body, Community Justice Scotland. This followed a process of consultation beginning in 2012, and is only the latest in a series of highly contested reforms. These reforms notably did not target the actual 'frontline' delivery of community penalties, which remained within local authority social work departments.
The research reported here is the first study of these reforms, and the most up-to-date work on the organisation of Scottish community justice. It advances the study of Scottish penal policy particularly as it relates to community penalties, drawing on work by McNeill (2005McNeill ( , 2019a as well as Morrison (2015) and Buchan and Morrison (2018). The theoretical focus is on two closely related Bourdieusian concepts -the penal field and the agonistic framework -which emphasise the role of (often hidden) conflict in the development of penal policy and practice. The penal field is adapted from Bourdieu's social theory and developed in Page's (2011) book, The Toughest Beat. This space is where penal change happens -'the social space in which agents struggle to accumulate and employ penal capital-that is, the legitimate authority to determine penal policies and priorities' (Page, 2011: 10).
Work co-authored by Page (Goodman et al., 2015(Goodman et al., , 2017 brings to the foreground the role of struggles, including hidden struggles, in driving change in the penal field. The 'agonistic framework' they use to analyse changing orientations in imprisonment in California, and latterly the wider US, is founded on three 'axioms': • • 'Penal development is the product of struggle between actors with different types and amounts of power' • • 'Contestation over how (and who) to punish is constant; consensus over penal orientations is illusory.' • • 'Large-scale trends in the economy, politics, social sentiments, inter-group relations, demographics and crime affect (or condition) -but do not determinestruggles over punishment and, ultimately, penal outcomes.' (Goodman et al., 2017: 8, 13) In Breaking the Pendulum, Goodman et al. (2017: 123 n1) suggest using the framework in other jurisdictions and areas of criminal justice practice and policy. Recent criminological research, including a symposium published in Law and Social Inquiry (44(3)), has taken up this suggestion -deepening, qualifying and enriching the theoretical value of the agonistic framework. This article contributes to the nascent 'agonistic' literature by applying the framework to the community justice restructuring, potentially opening the way for further agonistic work on recent reforms in small jurisdictions. It builds on McNeill's (2019a) use of the concept by looking at more recent restructuring, which has been focused more on organisation than on the practices of punishment. It also refines the theory by mobilising Koehler's (2019) argument about the limits of 'agonistic' penal debate, and combining it with original insights about circumscribed penal fields, to explain a case characterised both by agonism and by only very limited penal change. Additionally, it contributes to a growing scholarship critical of exceptionalist narratives about Scotland's 'welfarist' penal policy (Brangan, 2019;McAra, 2017;McCulloch and Smith, 2017;McNeill, 2019b).

Outline
For much of this article, as in the reforms themselves, the term 'community justice' is used. I use the Scottish Government (2014: 1) definition of 'community justice' as '[t]he collection of agencies and services in Scotland that individually and in partnership work to manage offenders, prevent offending and reduce reoffending and the harm that it causes, to promote social inclusion, citizenship and desistance'. Community justice therefore entails a range of organisations and structures and services beyond the 'frontline' delivery of penal supervision in the community (which would be implied by 'probation' in many jurisdictions and 'criminal justice social work' (CJSW) in Scotland).
In the first part of the article, I use the agonistic framework to sketch the contours of hidden conflict -particularly between localist and centralising interests -in recent reforms to the system of Scottish community justice. The second section contrasts this with a broad policy consensus in Scotland over the need to reduce the country's rate of imprisonment, and poses the question of how these contradictory features can be reconciled. The final section draws on recent developments in the literature on the agonistic framework, particularly Koehler's (2019) argument that penal fields are subject not only to 'agonistic' debate but also to 'antagonisms' -agreement between penal policy actors over the exclusion of some positions from the sphere of debate. The article combines this with insights about the circumscription of penal fields in smaller jurisdictions, arguing that these refinements to the agonistic framework render it able to explain the paradox of the failure of decarceration in Scotland despite widespread consensus at the policy level.

Restructuring community justice in Scotland
After briefly describing Scotland's penal field, this section uses the agonistic framework to examine a specific area of Scotland's penal policy: the restructuring of community justice under the 2016 Community Justice (Scotland) Act. It argues that this policy change has been shaped by hidden conflict, particularly the struggle between advocates of centralising the system and those in favour of local delivery, in line with the principles of the agonistic framework.
Although the focus of this article is theoretical and historical, this section also draws on empirical data: qualitative semi-structured interviews with 21 practitioners and policy actors from various parts of Scotland's penal field, which took place between August 2014 and January 2015 (Buchan, 2017). The sample included seven women and 14 men, whose length of service varied from a few years to several decades. The interviewees included staff from CJAs, CJSW employees, third sector staff and politicians, representing a range of perspectives on the then-imminent reforms.

The penal field in Scotland
Community justice is conceived as a 'subfield' of Scotland's penal field, which also includes institutions such as the Scottish Prison Service (SPS), the police, and courts service. Under a Scottish National Party (SNP) government agenda driven by imperatives of asserting Scottish distinctiveness and legislative competence (MacLennan, 2016), major change has taken place across these institutions. This includes a turn towards a desistance agenda in SPS (McConnell et al., 2013), the merging of police forces into a single national force in 2013 (Fyfe et al., 2018), and the geographic restructuring of the courts service.
Community justice has held an unusual position in Scotland's penal field since at least the 1968 Social Work (Scotland) Act, which abolished Scotland's probation service and merged its functions with local authority social work departments. CJSW is hence both a criminal justice service and a local authority social welfare agency. Because of the complex and overlapping nature of criminogenic needs among people supervised in the community, and the longstanding role of the third sector, here as elsewhere (McWilliams, 1983), this subfield includes a range of actors from the public and third sectors, not primarily or traditionally concerned with punishment but nonetheless involved to some degree with community penalties.

Hidden conflict
Central to the agonistic framework, and underpinning two of its axioms, is the claim that there is almost constant hidden conflict even during times of apparent consensus (Goodman et al., 2015). Subsequently, Goodman et al. (2017) draw on the penal field's focus on hidden actors and power struggles, and its breadth, to highlight the roles played by conflict and by non-justice actors in shaping imprisonment in several parts of the US. They also advise activists and scholars to attend to periods of (apparent) consensus and stability as well as those of conflict and change.
The Scottish case provides an opportunity to apply the agonistic framework in a different context. Drawing on Kelly's (2017) study of Scottish youth probation, as well as his own historical research, McNeill (2019a) highlights conflicts over the use of volunteer as opposed to police probation officers, and over probation practice leading up to the 1968 Social Work (Scotland) Act. In this case, conflict was hidden by limited historical evidence, largely from official sources, giving an impression of consensus (City of Glasgow, 1955;McNeill, 2005). The comparatively 'hidden' nature of probation and community penalties in relation to the rest of the criminal justice system (Robinson, 2016) may also have been a factor limiting the historical evidence for conflict available to criminologists in Scotland. This section builds on McNeill's (2019a) work applying the agonistic framework to Scotland, but brings this approach up to the present day. It is also more focused on administrative structures than on probation practice. Central andlocal, 1989-2007 Although concerns over local variation and inconsistency long predated the 1968 Act (Kelly, 2017), it was one of two key events that brought the central-local debate to primary importance in community justice. It cemented community penalties within local government and a 'generic' social work system, with longstanding cultural importance to narratives of Scottish penal distinctiveness (Brodie et al., 2008).
The second such event was the shift by the Scottish Office (the section of UK government with responsibility for Scottish affairs pre-devolution) to ring-fenced, centralised funding for criminal justice social work in 1989. This was accompanied by National Objectives and Standards (Rifkind, 1989) to which social work had to adhere. Local authorities had traded some autonomy and power for financial security; a further result was the partial re-specialisation of social work with offenders and the emergence of the term 'criminal justice social work' (CJSW).
The ring-fencing initiated contested restructuring through the 1990s and 2000s (Morrison, 2015), in the wider context of profound political changes -the reorganisation of Scottish local government from regions into 32 unitary authorities, Scottish devolution and the (re-)establishment of Scotland's parliament (with wider criminal justice implications discussed elsewhere: McAra, 2005). Repeatedly, central government (particularly post-devolution) attempted to centralise community justice to improve efficiency, place CJSW under stricter control and bring about closer integration with other justice agencies. These attempts were resisted (with varying success) by 'localist' interests, who emphasised the importance of responsive local delivery.
These 'localists' comprised not only local authorities, but also professional social work bodies -an alliance created by the 1968 Act (Brodie et al., 2008). That legislation continued to exercise an effect through the discursive formation around 'generic' social work -localists could portray local delivery as vital to maintaining the 'social work values' essential to Scotland's distinctive penal welfarism (CoSLA, 1998). The conflict only emerged briefly into the 'open' in the mid-2000s when Scottish Labour's (2003) manifesto proposed merging SPS and CJSW into a single correctional service. This provoked fury among local authority and social work interests (Denholm, 2003) and concern among academics and practitioners that this posed a threat to CJSW's distinctive welfare-oriented identity and culture (McNeill, 2005).
This reaction, and the moderating influence of the Liberal Democrats as coalition partners, produced a further compromise -the establishment of the CJAs in 2005. These eight regional bodies would have responsibility for governance, oversight and commissioning of community justice services within their areas. At the time, CJAs were seen as resolving the tensions between local and central control, but they too would soon be superseded (Buchan and Morrison, 2018). Conflict between central government and 'localist' interests (including local authorities and social work bodies) shaped the structural compromises around community justice governance in Scotland in 1989Scotland in -2007; a focus on hidden conflict could therefore be helpful to analyse the most recent reforms.

A fourth compromise: The 2016 reforms
In 2012, two reports emerged which initiated the latest reforms, sealing the fate of CJAs. The report of the Commission of Women Offenders (2012), chaired by Elish Angiolini, touched only briefly on community justice but in doing so highlighted several specific problems with it -short-term and inconsistent funding, a lack of accountability and leadership, and a 'cluttered landscape' of provision -and recommended replacing CJAs with a national service.
A subsequent report by the independent public auditing body Audit Scotland (2012) made no such recommendations but added further to the impetus for reform by exploring in depth the issues around funding and accountability. There began then a process of consultation over what would replace the CJA system, which lasted several years and ultimately led to the replacement of CJAs with a new two-tier system comprising local partnerships and a national leadership body.
The consultation process for the restructuring began very shortly after the reports were published. The Scottish Government suggested three possible structures -fully local provision, a national service and an 'enhanced CJA' model. The interview data and consultations show the different options attracted support from conflicting interests.
The CJA staff expressed a preference for the 'enhanced CJA' model, but their influence was limited. Even if CJA staff had been vocal, they could not have a significant role as there were only ever 24 full-time equivalent staff spread across eight CJAs in Scotland (compared to several hundred CJSW employees), and their powers were circumscribed -both formally, and informally by the need to maintain good working relations with local authorities (Buchan and Morrison, 2018). Nonetheless, a theme that emerged in the interviews with CJA staff was a sense of frustration that CJAs would be abolished just as they were beginning to contribute in a role of their own, albeit not the one intended for them (Buchan and Morrison, 2018): I think at one point we thought we would be working with the government to look at how we could enhance CJAs. That seemed to be more the way Audit Scotland were going, you know, let's look at . . . They didn't really make recommendations, but reading between the lines, I thought they thought we could be improved . . . But when it became quite clear that the redesign was really going to be a local versus national argument . . . (CJA Chief Officer) National politicians and others involved in central government policymaking tended to favour centralisation, seeing the value of promoting consistency across Scotland and avoiding reduplication of effort: Some of this comes about because of the dysfunctionality of local government in Scotland, who basically oppose centralisation even when it's not about centralisation.
I'm a great believer in decentralisation and localism . . . but I agreed with the analysis that Elish Angiolini had that there was not a champion for criminal justice at a national level (MSPs [Members of the Scottish Parliament]) Local authority social workers and CJA elected members (local councillors serving as board members) favoured a local approach to organising community justice, which was seen as the only way to make services responsive to local needs in the context of geographic and social diversity and contrast between Scotland's local authority areas: [I]n [council area] we have very farming and rural things, so it's expensive things that are getting damaged -tractors, these quadbike things that farmers use for going round the hills and stuff -and that's particularly prevalent in the [rural area], and somewhere in [city] things like that doesn't happen, it's maybe like car thefts and drunk and disorderly kind of offences and opportunistic stuff . . . (CJA elected member) After six months, consensus had not been reached, although both local and national approaches had attracted support. The Scottish Government therefore negotiated with local government bodies to develop a compromise 'Option D' (Scottish Government, 2014). Under Option D, a new set of local partnerships, aligned with the community planning framework, would take on most of CJAs' core responsibilities for planning, commissioning and evaluation within their local areas. At the same time, a new national organisation, Community Justice Scotland, would be established to oversee and support local provision while acting as a 'champion' for community justice. The new body would not have the power to hold local authorities or local partnerships to account for poor performanceavoiding the 'tangled' lines of accountability and attendant local authority distrust that had bedevilled the CJAs (Audit Scotland, 2012;Buchan and Morrison, 2018), but stopping some way short of what Angiolini had recommended. There is currently still scope for government to extend the central control of the system -the 2016 Act (section 4(5)(b)) provides for the Scottish Government to transfer functions and powers to Community Justice Scotland at a later time, although doing so would likely attract scrutiny.
'Option D' was the third time since 1989 that the Scottish government's efforts to centralise community justice had fuelled conflict with local government. In every case, neither of the deeply entrenched positions could prevail fully, leading to a compromise between local and national interests and hence to community justice institutions that were neither 'fully' central nor local.
Approaching the community justice reforms with a focus on hidden conflict brings to light the disagreements between 'localist' and centralising interests to explain the pattern of frequent restructuring.

Conflict at the edges
An agonistic approach to the consultation process leading up to the reforms also highlights the question of which actors can participate in these conflicts. After successive restructurings, and amid ongoing efforts to develop partnership structures (Commission on the Future Delivery of Public Services, 2011), the community justice field in Scotland -already liminal to the penal field as a whole, with limited autonomy (Krause, 2018: 6) -is diffuse enough to raise the question 'Where does this subfield begin and end?' The fact that there is sometimes struggle over the boundaries of the penal field does not indicate how to determine where these boundaries are. This is also a problem with Bourdieu's field theory (Krause, 2018), but is highlighted when investigating such a diffuse and complex area of criminal justice.
The boundaries of the penal field may themselves be a site of struggle between actors (Bourdieu, 1983: 324); the controversy over the 2003 single agency proposals is a particularly prominent example of CJSW (and allies) defending the boundaries and autonomy of community justice against the encroachment of their better-resourced, more powerful rival SPS.
The process of consultation leading up to the 2016 reforms can be understood as conflicts over the boundaries of Scotland's penal field. A theme that emerged from discussions of the consultation was the predominance of local authority social work in written responses and consultation events, particularly in the first round. One charity executive described how local government's greater financial resources limited the third sector's ability to play a role: [I]f you look at the number of people that attended those consultation groups, they were packed by local authority. They had a three-line whip. Every community justice social worker was out there. Every social work manager, everybody that they could mobilise. I would think on average -I went to three of them, and there would have been no more than two or three people from any other sector than community justice social work in any of those consultation meetings. And if you look at the responses you'll find the same sort of thing . . . absolutely flooded by community justice social work [sic] and local authorities saying "what we want is more power to community justice social work"! Ha, really? . . . I mean organisations like ourselves, we can't send . . . we can't afford to send people. I haven't got staff to send to these things. (Third sector manager) This is in line with critical perspectives that suggest that in the context of criminal justice contracting, charities have tended to have 'a seat but not a voice' (Crisp et al., 2011).
The consultation and subsequent decision-making processes were notable for their length -beginning in November 2012 and not fully concluding until the closure of the CJAs in April 2017. CJAs spent half their life operating with the threat, and then the imminent reality, of abolition. Although not aimed at 'frontline' CJSW practice, the reforms did affect it -practitioners interviewed in the study described the reforms as disruptive and time-consuming even during consultation, as well as creating uncertainty for CJA employees.
Given the compromised basis of the latest structure, it is likely that the 'long struggle' (Goodman et al., 2017) over the structure of Scottish community justice will continue. Indeed, there are signs that further structural reforms may be forthcoming, with the recent establishment of a national steering group to strengthen community justice in Scotland (Davidson, 2019). Although there has been extensive hidden conflict over how the system should be organised -and over who should have a voice in the process -the Scottish Government itself admits that there is no direct relationship between the organisational structures surrounding community penalties in Scotland, and the rates of reoffending and imprisonment (Scottish Government, 2012: 5).
The process of consultation leading up to the 2016 Act was characterised not only by conflict over the question of local or central delivery, but also over the penal field's boundaries and which actors could participate in it, over a long and disruptive period of policy development. Dynamics of inclusion and exclusion in penal fields are also relevant to a major question facing Scottish criminal justice -the lack of progress in reducing prisoner numbers. The next section discusses this, before the article returns to the boundaries of penal fields.

Policy consensus and decarceration in Scotland
There has been significant hidden conflict over the structure of community justice in Scotland. However, the drive to repeatedly reform the system has been the result of a high degree of policy consensus, over the need to reduce Scotland's high prison population by diverting cases from prison to community penalties. Ostensibly this is a welfarist, 'evidence-based' and distinctively Scottish approach to justice (McAra, 2017), despite its failure to effect decarceration, discussed below.
Penal fields have a distinct orientation at any given time, which defines the goals and means of penal action. Goodman et al. (2015) use the agonistic framework to explain the change in orientations in California's penal field, from treatment and rehabilitation to incapacitative imprisonment. This section will show that the prevailing orientation of Scotland's penal field is towards decarceration through diversion to community penalties, even if that does not necessarily guide all action within the penal field (just as the turn towards incapacitation in Californian imprisonment did not eliminate all rehabilitative programmes). It will also consider how this consensus is connected to narratives about Scotland's penal identity, constructed as being distinct from (and better than) that of England and Wales.
Particularly important in this regard was Scotland's Choice, the report of the Scottish Prisons Commission (2008). The titular dilemma was between two possible penal futures -to retain high imprisonment and reoffending, at immense economic and social cost, or (as the Commission recommended) to shift to a model in which prison is a true last resort and community penalties are the default. The nationalist Scottish National Party (SNP) had been elected for the first time the previous year on a manifesto which included a reduction in short prison sentences. Helping to set the tone for SNP justice policy was the explicitly comparative aspect of the Scotland's Choice recommendations -Scotland could follow the path of England, or those taken by Ireland or Scandinavian nations.
Hence, the effort to reduce imprisonment (as the prison population in England and Wales rose sharply until 2011, amid 'tough on crime' political language) is deeply connected to a post-devolutionary impulse towards nation-building and establishing a distinct political identity for Scotland. This impulse had informed a process of criminal justice 'hyper-institutionalisation' in the early years of the Scottish Government (McAra, 2008), but under the SNP -in its third successive term in office, at the time of writing -it is also connected to establishing Scotland's position distinct from England in matters of penal and social policy (Mooney et al., 2015).
The first legislative step in this agenda was the 2010 Criminal Justice and Licensing (Scotland) Act that introduced Community Payback Orders and a presumption against short sentences (PASS) in favour of community penalties, as well as establishing the Scottish Sentencing Council. The 2010 Act attracted significant controversy in Parliament, with Labour and Conservative members largely opposing PASS particularly because of concerns over knife crime (which had been a focus of punitive proimprisonment legislation under the preceding Labour government), and over the use of short sentences as 'respite' for communities from the most persistent offenders, which was also a concern for members of the judiciary who criticised the plans. These controversies were widely reported in Scottish news media (Howie, 2009). Compromise with the SNP's Liberal Democrat coalition partners, and with judicial interests, required the presumption to be reduced from sentences of six months to three months (Hutton and Tata, 2010).
Decarceration through diversion to community justice is politically useful to the SNP as a way of reasserting Scotland's distinctiveness in penal matters (Maclennan, 2016;Mooney et al., 2015), and it is not surprising that it should remain government policy while the SNP have held power through successive elections (including an overall majority in 2011-2016). However, this approach has increasingly attracted cross-party support in Scotland. Parliamentary debates leading up to the 2016 Community Justice (Scotland) Act (Scottish Parliament, 2015) demonstrated broad support for the principle and particularly over the need to reduce short prison sentences. In the 2016 Holyrood election, even the Scottish Conservatives' (2016) manifesto focused on making community sentences "credible alternatives to prison". Reducing imprisonment, including by extending the presumption, is a policy commitment in the Programme for Government (Scottish Government, 2019a); under the banner of 'smart justice' it is the stated objective of the new body Community Justice Scotland (2017).
The consensus is sufficient to suggest that the 'orientation' of Scotland's penal field is towards decarceration through diversion of cases away from prison (especially short periods of imprisonment) and into community penalties. However, the existence of a particular orientation does not necessarily mean that its goals will be achieved.

The failure of decarceration
Investigating the dynamics of community justice reform, in the context of Scotland's efforts to decarcerate, reveals a complex and contradictory situation. There is extensive hidden conflict over several issues, particularly the administrative structure of community justice and whether it should be centrally or locally organised. However, there is also substantive consensus over the value of community penalties over imprisonment and the need to reduce imprisonment in Scotland. The existence of policy consensus around decarceration does not contradict the argument of the agonistic framework, which allows for consensus as well as conflict (Rubin and Phelps, 2017;Page et al., 2019) (which in any case can only be defined in relation to each other). It does however highlight a limitation of the agonistic framework -it does not give any indication about which conflicts will shape penal change, or how.
In the decade following Scotland's Choice (Scottish Prisons Commission, 2008), there has been little progress in reducing imprisonment (McCulloch and Smith, 2017). The Scottish prison population peaked slightly in 2011-2012 before slowly falling to 2007-2008 levels by 2018. During this time, community sentences increased their 'market share' from 14% to 19% of criminal cases but -here as in other jurisdictions (Phelps, 2013) -this gain was mainly at the expense of fines, not imprisonment (McNeil, 2019b;Scottish Government, 2019a), suggesting substantial up-tariffing taking place even during a fall in recorded crime. Between 2007Between -2008Between and 2017Between -2018, the 'share' of sentences occupied by imprisonment actually increased slightly (from 13% to 14%).
It is also noteworthy that, while a brief mention of CJAs by the Commission on Women Offenders' (2012) was enough to trigger the 2016 reforms, progress towards its other recommendations has been slower. The recommendation of a national community justice service was accompanied by several others about what sorts of service it would provide, including a recommendation that: Community Justice Centres (one stop shops based on the 218 Service, Willow Project and Women's Centres in England) are established for women offenders to enable them to access a consistent range of services to reduce reoffending and bring about behavioural change.
(Commission on Women Offenders, 2012: 7) A related development was the last-minute decision to reverse the construction of a new women's prison at Inverclyde, with the intention of building several local Community Custody Units instead. These would be small (in line with Scandinavian evidence on effective prisons - Johnsen et al., 2011) and focused on maintaining links between prisoners and their communities. Years later, there has been almost no progress towards either of these outcomes.
If decarceration in 2008-2018 was limited and halting, the past year has seen it go into reverse with troubling speed. At the time of writing the Scottish prison population is at an all-time high of over 8200 (an imprisonment rate of around 155 per 100,000 -SPS, 2019). This appears to have been driven largely by an increase in average prison sentences as well as a decline in certain forms of supervised release, notably the Home Detention Curfew, following a murder by a prisoner who had absconded (Yousaf, 2018). This has occurred despite a broad political and practitioner consensus on decarceration, supported by a vibrant and effective community of criminologists and penologists who have enjoyed relatively open access to policymakers and politicians in Scotland, especially compared with England and Wales (Morrison and Sparks, 2015).
Although at odds with the aforementioned diversionary consensus, this situation chimes with Cohen's (1985) 'net-widening' thesis -that community penalties serve to draw more people into the criminal justice system, serving as supplements rather than alternatives to prison. This thesis is also supported by recent evidence from across Europe where increasing use of community sanctions has mostly failed to reduce prison populations (Aebi et al., 2015).
However, net-widening is not inevitable; Phelps (2013) shows that probation can be truly diversionary under the right sentencing and probation practice conditions. The current mix of conflict over structures and policy consensus has not produced substantial progress towards decarceration, because neither has actually addressed these conditions. The reasons for this can be found in a refined version of the agonistic framework, which is discussed in the next section.

The limits of agonism
This article has shown that the agonistic framework can help to explain recent structural changes to the system of community penalties in Scotland, by drawing our attention to hidden conflicts between different actors within the penal field (primarily, in this case, local authority/social work interest and central government). However, it has also established that there is broad consensus in the Scottish penal field, not only over ends (the reduction of imprisonment) but also over the primary means (the diversion of cases towards community penalties). Despite the importance of consensus in shaping the orientations of penal fields (Page, 2011), progress towards decarceration has been slow in Scotland, even though a consensus towards it exists (and significant conflict has taken place over the administrative detail of how it is to be achieved).
How can the agonistic framework help us to understand and explain this situation? Recent work by Koehler (2019) adds a valuable refinement to Goodman et al.'s (2015Goodman et al.'s ( , 2017 theoretical approach, responding to their focus on conflict and relative inattention to stability. Koehler suggests that while any penal field will be characterised by agonisms -conflicts over matters generally regarded as legitimate -it also has what Koehler calls 'antagonisms', positions of consensus around matters widely accepted not to be matters for legitimate debate. Agonisms may include debates over the purposes of imprisonment, or the extent to which it should be used -but actors within a penal field are unified against positions they consider illegitimate, strange and unworthy of serious consideration (Koehler uses the example of crucifixion as criminal punishment). The boundaries of the penal field, and the limits of debate, are constructed between agonisms and the 'antagonisms' in which actors are unified against especially unworkable or strange positions.
The finity of penal fields I combine this insight with a new one, evident but not always acknowledged -penal fields are not infinite, because the public and policy sphere has only limited space for arguments and debates. Time for Scottish Parliamentary and Committee debates on policy and legislation is limited, as is the number of person-hours available to relevant government institutions such as the Justice Directorate of the Scottish Government. There is also a limited number of criminal justice professionals, and their time to engage with policy debates is also limited and circumscribed, particularly in the context of rising caseloads and austerity-driven pressures towards efficiency (McNeill, 2019b).
Hypothetically, if all other social and political conditions could somehow remain identical, it would follow that the penal field of a smaller place (like Scotland) would be smaller, with less room for debate, than that of a larger place (like California or the entire US). There are fewer criminal justice employees and policy actors at all stages of the criminal justice policy process; for example, the California Department of Justice (2020) employs 4500 people, nearly as many as the entire Scottish Government (2020). The California Department of Corrections and Rehabilitation (2020) boasts a plethora of executive staff while SPS has only three Executive Directors and a single Chief Executive.
Of course, no two jurisdictions or nations do have the exact same set of social and political conditions, and it can therefore be expected that these differences would produce penal fields of differing sizes in nations of a similar population.
The question of finity therefore has implications for comparative research on punishment, particularly when comparing nations with penal fields of differing relative size. For instance, while Japan is a populous nation, its penal field is tightly circumscribed, with a high degree of control exerted by a small elite of senior prosecutors and judges (Fenwick, 2013). This may explain the slow pace of penal change and dominance of 'crime control values' in the Japanese system. Conversely, deliberative political cultures and coalitionoriented democracies have been widely credited for low imprisonment in European social democracies, including cases such as Finland where this has required significant policy change (e.g. Lappi-Seppälä, 2011). Space does not permit a detailed analysis of the size of different penal fields, which would need to take the form of detailed and sensitive qualitative research dealing carefully with the question of 'edges' discussed above and with which institutions may be counted as part of the penal field. As Barker (2006) argues, traditions of democratic involvement influence the range of organisations and perspectives which can enter the penal field; more deliberative approaches allow a wider range of viewpoints to be represented (McCulloch and Smith, 2017). Different resources and power allow actors to take up more 'space' in the penal field and/or to shape how that space is used. Time is also a limiting factor on the 'size' of the penal field -and some actors (such as local government social work) can devote more time than others (such as charities) to penal struggles, as in the case of the community justice consultations where charities were less able to have a voice.
Therefore, even positions that might attract considerable sympathy from criminal justice elites (such as drugs decriminalisation) can be 'crowded out' of the penal field by agonisms that take up more space and time. Perhaps most obviously, penal fields tend to be dominated by questions about how best to punish rather than whether we should punish.
In applying this to community justice restructuring, I draw on Nellis' (2016) criticism of the 2016 community justice reforms and described CJSW as the "weak link in Scotland's penal reimagining". This "reimagining" included Scotland's Choice and the decision not to build the new women's prison at Inverclyde. Nellis argues that discussions surrounding the 2016 Act had become dominated by administrative debates over whether the system should be organised centrally or locally, and as a result: While tremendously well skilled at face-to-face work, [CJSW] desperately needs to reimagine itself as an organisation, a profession, committed to the reduction of the prison population and to the creative forms of practice, including but not privileging more integrated uses of EM, necessary to achieve it. (Nellis, 2016: 22) There is a similar problem of 'crowding out' of more radical debates and questions, at the larger scale of the entire Scottish penal field, and an agonistic approach (as modified above) highlights the specific dynamics of conflict and consensus that produce this situation. Underpinning the post-2008 decarceration consensus is an assumption that if 'credible' community penalties are available and delivered by an organisationally optimised (after several tries!) system, then sentencers will decide to use them instead of imprisonment and the prison population will fall. This evinces a consumerist logic of criminal sentencing, with the sentencer as 'consumer' with the sovereign right to choose between various options. Tata (2018), focusing on pre-sentencing reports, argues that this logic places the onus on criminal justice social workers (who cannot influence sentencing decisions) to persuade the sentencer towards a community penalty, and away from prison -thereby reinforcing the idea that prison is the 'backstop' to any criminal sentence. Hence, a consumer logic undermines efforts towards decarceration, making sentencers less likely to impose a community sentence in place of prison. In any case, there is no evidence that the organisational features of the pre-2016 system were undermining judicial confidence in community penalties or making them less likely to choose these sentences.
The debate in Scotland's penal field has focused on the best way to use community penalties as a release valve for imprisonment, and has failed to address whether or not using community penalties in this way is practicable or ethical. It is not working -as McNeill (2019b) has shown, community penalties have grown in Scotland at enormous speed, despite a fall in crime and without reducing imprisonment. Given the recent rise in the prison population, Scotland now has the highest combined prison and community supervision rate in Western Europe (Aebi and Hashimoto, 2018).
In any case, much of the increase in imprisonment is accounted for at the other end of the severity scale, by increases in length of prison sentences (including minimum tariffs on life sentences) (Scottish Government, 2019) rather than the decision to use imprisonment over community penalties in 'borderline' cases. Similarly, 'crowded out' of the debate is the Scottish Prisons Commission's (2008) recommendation of capping the prison population at 5000.
Applying the agonistic framework to the process of community justice restructuring in Scotland highlights the importance of hidden processes of conflict and negotiation in these reforms. In particular, conflicts between centralising and localist tendencies have shaped the institutional structure of community justice and will almost certainly continue to do so. The consultation events themselves were also sites of conflict over which actors were able to influence the process. Koehler's (2019) insights about the construction of the limits of debate, and the finity of any penal field, help to explain why despite broad consensus around its importance, and the structural reforms underpinned by that consensus, the project of decarceration in Scotland has made only very limited progress.

Concluding remarks
This brief section concludes the article by considering some possible future avenues for research. While the agonistic framework was developed to analyse penal change in large polities (California and the USA), this article has developed its ability to explain recent penal change and inertia in small jurisdictions such as Scotland. By applying the agonistic framework to the case of community justice reform and decarceration policy in Scotland, it has shown that a version of this theory -refined both by Koehler's (2019) and my own insights -can be employed to analyse patterns of conflict and consensus over penal policy, and particularly to explain why these do not always produce substantive change even where there is broad consensus over policy goals. Scotland emerges as a jurisdiction whose relatively circumscribed penal field enabled the 'crowding out' of more radical positions in favour of a narrow and questionably effective focus on the promotion of community sentences as diversionary measures.
These insights may be particularly relevant for the analysis of criminal justice policy in other small European jurisdictions -including Ireland and the Scandinavian nations, whose lower imprisonment rates were invoked by the Scottish Prisons Commission (2008). Further investigation in this area could also compare penal fields to analyse the extent to which their size (in comparison with other factors) affects the capacity for penal change. Research in Scotland, meanwhile, might profitably investigate the performance of the new community justice structures.
It is notable that Breaking the Pendulum itself, and subsequent work that builds on its theoretical contribution (McNeill, 2019a;Koehler, 2019;Page et al., 2019;Rubin and Phelps, 2017), are focused on western liberal democracies -in which policies are formed largely through somewhat open debate between a range of actors, in the context of elected governing parties and strong civil service institutions. A fruitful -albeit challenging -future direction for the agonistic framework might be to consider the dynamics of penal change in jurisdictions where the freedom to debate policy is limited, and where the limits of what is legitimate may be coercively enforced. Such an investigation might shed light on how closely the agonistic framework and western-style democracy are connected.