Introduction
Contemporary British counter-terrorist legislation is dominated by ‘counter-law’: using law against law, it systematically undermines the rule of law. This proposition will be substantiated by developing a detailed ‘archetypal’ account of the rule of law considered as a critical ideal. The rule of law will be identified with four tendencies in law – towards greater universality, knowability, followability and justifiability – and ‘counter-law’ with tendencies to block or reverse all of these. Counter-law tendencies in contemporary counter-terrorist legislation will be discussed in detail. This critique will also be related to contemporary debates on law and counter-law; it will be argued that critiques which relativise or historicise the liberal model of the rule of law fall short by failing to engage with it on its own terms, thereby undervaluing its utility as a normative resource.
The first section engages with contemporary scholarship on threats to the rule of law; it notes that many authors relativise or problematise the liberal legal order, returning it to its historical context or focusing on its contradictions and blind spots, and advocates a fuller critical engagement with the rule of law. The second section expands on Richard Ericson's concept of counter-law by developing a model of the rule of law as a critical ideal. Lon Fuller's criteria for legal systems are synthesised to give four general demands: laws should be universal, knowable, followable and justifiable. The status of the rule of law and its elements as ideal-types is discussed, and Nigel Simmonds’ ‘archetype’ model is found to have greater normative and analytical power than Matthew Kramer's ‘threshold’ model. The third section discusses terrorist offences, with particular reference to inchoate, preparatory and situational offences, while the fourth discusses the definition of terrorism in English law. It is argued that terrorist offences are structured, not to allow subjects to conform their own behaviour to the law, but to facilitate the pre-emptive interruption of loosely-defined patterns of behaviour judged by experts to be undesirable. The fifth section discusses the range, significance and gravity of the departures from the rule of law that have been identified, considering some counter-arguments before drawing conclusions for policy-makers and legal scholars.
Legality and its critics: a call for rethinking
A key concern in recent legal scholarship has been the articulation and defence of a liberal model of law, criminal law in particular. This model has been counterposed to developments perceived to threaten core liberal legal principles such as the presumption of innocence.Footnote 1 It is argued that, in the last two decades, there has been significant movement away from a realised liberal legal order – characterised by universal access to justice, practical equality of legal rights, effective procedural justice and retroactive sanctioning with at least some reference to desert – in the United Kingdom and perhaps more widely; and that this is a negative development, to be resisted or at least deplored. Substantial literatures have built up around broader trends within this movement: changing practices and principles of criminalisation;Footnote 2 the challenges to the legal order posed by society's responses to anti-social behaviour and ‘incivilities’,Footnote 3 on one hand, and terrorism and extremismFootnote 4 on the other; the innovations grouped under the heading of ‘preventive justice’;Footnote 5 and the executive-led undermining of legality by a countervailing trend of ‘counter-law’Footnote 6 or the parallel development of ‘enemy criminal law’.Footnote 7
As this broader literature has developed, the liberal legal model has itself come under criticism. Five main lines of argument have been advanced. The historicist argumentFootnote 8 situates the liberal legal model within its – relatively brief – historical moment: by dethroning the model as we currently understand it from the heaven of concepts,Footnote 9 historicist critics aim to call into question both the taken-for-granted virtues of the model and the assumption that any deviation from it must be challenged as such (although these developments may be challenged on other normative groundsFootnote 10). Second, a dialectical argument,Footnote 11 building on the historicist approach, grounds the liberal legal model in a particular socio-political conjuncture and argues that the liberalism of the model is internally incoherent, so that it is incapable of being universalised. According to this argument, the liberal model harbours its own contradiction within it, and is ultimately inconceivable without it; what are perceived as deviations from the model may express the model's own inherent tendencies or fallibilities.
The third and fourth are variants on a pluralist argument, which concedes that the liberal legal model may provide an appropriate normative framework for the workings of the law, but stresses that the law is only one of several, more or less interdependent, regulatory systems. The descriptive variant of this argumentFootnote 12 asserts that social order in advanced societies is produced both through the law and through regulatory systems not mediated through law. A critical variant focuses on policing in particular, arguing that policing as a domain of regulation is ‘unconstrained not only by principles of law, and of legitimacy, but by principles of any kind’.Footnote 13 Lastly, a more sceptical regulationist argument has it that law is, simply, one regulatory system among others, to be judged only by its effectiveness in producing sustainable social order: the regulationist has no investment in the liberal legal model, or any other normative model.Footnote 14
These arguments can be seen to overlap, and in practice critiques of the liberal model may draw on two or three of them. However, it is not clear that all five are consistent with one another; in particular, the dialectical and critical pluralist arguments engage (however critically) with a normativity which is bracketed or ignored by the other three. Moreover, these arguments make only a qualified challenge to the normative standing of the liberal model. The regulationist argument merely sidesteps it, establishing that other criteria of effectiveness are available. The dialectical and critical pluralist arguments problematise the model by arguing that it and its opposite are mutually implicated, whether its opposite is considered as internal to it or as a ‘police’ model existing alongside it. These arguments call for awareness of the areas that the putatively universal liberal order omits to cover, while also suggesting that such lacunae may be intrinsic to the model. Thus, to speak of ‘citizens’ who have ‘rights’ is always to invoke non-citizens who are denied those rights;Footnote 15 this recognition forecloses the possibility of celebrating citizenship rights as an absolute and unqualified good.Footnote 16 The historicist and descriptive pluralist arguments for their part relativise the normative model by localising it within a particular social and historical setting. To relativise the presumption of innocence, for example – establishing that it is broadly irrelevant in business regulation, and/or that it had no role in criminal justice until relatively recently –would be fatal to any argument grounded in the timeless and universal value of that principle.
These arguments have limited purchase on the model's normative value, however. Demanding rights for non-citizens does not preclude protesting at encroachments on rights that have been accorded to citizens, including through the affirmation of their (admittedly circumscribed) universality. Similarly, where the presumption of innocence does obtain, resisting its erosion is not necessarily mere conservatism or parochialism. Where norms have been used to guide practice and constrain the executive, those norms retain their power as critical standards which can be asserted against lawless power. Generally, the assumption that to make a normative claim is to assert an absolute – and hence that to relativise a principle is to deprive it of any claim to our normative consideration – is unwarranted. There is a risk of illusions of Olympian detachment, paradoxically, in the adoption of relativising perspectives: to name a value as having been current in debates within the community of legal scholars in the early twenty-first century is not, in itself, sufficient to remove oneself and one's audience from that setting. Norms recognised as such within a discursive community, particularly a community structured through continuing professionalised practice and reflection on practice, may serve as a critical resource for members of that community, even after reflection on the historical and functional specificity of their current usage.
The challenge for these critics is to relativise and problematise, not only the existence of norms of liberal legality, but the norms themselves, while giving their normative power its due. Given the capacity of these norms to constrain, as well as structure and legitimate, exercises of executive power, this stance is likely to entail at least a qualified defence of elements of those norms. This paper aims to contribute to such a critique, bringing Ericson's ‘counter-law’ model into dialogue with scholarship on the rule of law, and proposing a reading of the rule of law which situates it as a critical ideal.
Thresholds, archetypes and the problem of counter-law
Ericson introduces ‘counter-law’ by defining the law as ‘a liberal institution that holds accountable those who intentionally cause harm’;Footnote 17 elsewhere he defines the rule of law as the idea that individuals ‘should know what is and is not legally authorized … to ensure a predictable environment in which to make rational choices about rule-governed behavior’.Footnote 18 While the rule of law is associated with classical liberalism, Ericson argues, the rise of ‘neo-liberal social imaginaries’ has prioritised ‘the precautionary urge for certainty’;Footnote 19 this elevates pre-emptive harm prevention above the rule of law. The result is ‘counter-law’, which takes two forms: the enactment of ‘laws against law’ (counter-law I) and the creation of ‘surveillant assemblages’ (counter-law II). In both cases the goal is to pre-empt perceived threats by using the powers of the criminal justice system in coercive and incapacitatory ways. To make this possible, both forms of counter-law ‘erode or eliminate traditional principles, standards, and procedures of criminal law’.Footnote 20 Counter-law thus describes situations where legal powers are used to circumvent legal standards and principles, seen as obstacles to the safeguarding of power or public safety.
Conceiving counter-law as a tendency working against the realisation of the ideal of the rule of law enables us to define it more systematically. We can take as a starting-point Fuller's ‘morality of law’: the argument that laws, as well as existing, should be publicised, prospective, understandable, non-contradictory, compliable with, reasonably unchanging over time and congruent with their own official administration.Footnote 21 Fuller's criteria can be recast as three broad principles. First, laws should be equally applied to all: to say that the law in a certain jurisdiction requires one group of actions and prohibits another is to say that every individual within the jurisdiction is subject to these requirements and prohibitions, and no others. Secondly, laws should be knowable: it should be possible for an individual to establish what the law requires in a given situation, and to understand those requirements. This requirement of itself implies that the law should be comprehensive: there can be no area of social life to which the law does not apply and where it thus cannot be known. Since no set of rules can cover every possible eventuality, the law as a system must have comprehensible coherence, so that it can reasonably and consistently be interpreted and extended when necessary.
Thirdly, laws must be followable. Followability requires knowability: it is not possible to follow a law that one cannot identify or cannot understand. Further, it is not possible to follow a law that demands the impossible, or a law that has changed since one first tried to follow it. Followability also implies that it should be possible to choose to follow a law. This is an assumption about the subjects governed by law as well as about the law itself: laws in this understanding take the form of ‘general norms directed at, and capable of being understood and followed by, persons deemed to possess the necessary capacities’.Footnote 22 The subjects of law are conceived, crucially, as free: to be followable, laws may not structure social life to the point where no margin of choice remains. Nigel Simmonds argues that this association between followability and freedom to shape one's own life is fundamental to the rule of law: ‘wherever the rule of law exists to any extent at all, citizens will enjoy some zones of optional conduct that are protected from the interference of others’.Footnote 23 Thus we cannot say that we are living under the rule of law if our every action is determined by direction from others (even others who are enforcing formally valid laws), as in such a situation there is no undirected area of conduct left to be governed by law.
A fourth set of demands (not explicitly stressed by Fuller) follows from the second and third. Given that individuals choose to follow laws, it must be possible to justify a law: to give some explanation of a law which will motivate a rational hearer to follow it. However, reasonable people can disagree, and reasonable people (even in authority) can be mistaken: if a law can be justified it must also be possible for it to be rationally challenged, and for the challenge to succeed. As Jeremy Waldron argues, this criterion of justifiability – and the demand for procedural justice which it grounds – is itself intimately associated with the value of liberty: to present the rule of law merely as a system of clear and precise rules would be to ‘truncate what the Rule of Law[sic] rests upon: respect for the freedom and dignity of each person as an active center of intelligence’.Footnote 24 If laws are to govern free and presumptively rational citizens, they must be both rationally arguable and practically open to challenge.Footnote 25
In short, to say that a society exists under the rule of law is to make four claims about its laws. First, laws are universal: they apply equally and without exemption. Second, laws are knowable: it is possible to find out what laws exist and possible to understand them; there is no area of life outside or above the law; and laws coherently express general principles which make it possible to extend the law to cover new situations. Third, laws are followable: laws do not demand the impossible, nor do they remove any margin of choice. Fourth, laws are justifiable: it is possible both to explain laws and to make reasoned challenges to them, which have a chance of being effective.
The four demands sketched above do not constitute a coherent Weberian ideal-type, although any one of them alone might be developed into one. A system of laws that was perfectly universal might (in theory) be achieved by elaborating the law to levels of Borgesian comprehensiveness, with guidance for every detail of every individual's life. A system of laws that was perfectly knowable, by contrast, could be achieved by replacing the law with a handful of general precepts (‘As far as possible, do no harm’), which could be regularly advertised to the populace. A perfectly followable scheme of laws could be achieved by abolishing every offence which had ever been the subject of a contested trial, leaving only those prohibitions which society had shown itself to find unproblematic. A perfectly justifiable system, finally, would replace every sentencing decision with an open-ended debate of the type canvassed in Nils Christie's proposal for ‘neighbourhood courts’.Footnote 26
Whether these ideal-types, taken individually, are ideals in the normative sense is debatable. Certainly, as Matthew Kramer notes, the conjoined maximal realisation of each of Fuller's principles is not conceivable.Footnote 27 This question regarding the status of Fuller's criteria relates back to Hart's controversy with Fuller. Hart argued, with varying emphases, that ‘intelligent decisions’ could be taken with reference to some agreed ‘standard of criticism’ without the standard being a moral one;Footnote 28 that ‘the notion of efficiency for a purpose’ should not be confused with ‘those final judgments about activities and purposes with which morality in its various forms is concerned’;Footnote 29 and that moral status might be granted to criteria like those set out by Fuller, if it were granted that this morality was ‘compatible with very great iniquity’.Footnote 30 What is consistent throughout Hart's differing formulations is the argument that, while Fuller's criteria may describe the minimum prerequisites of a functioning legal order, they are in effect technical rather than moral virtues (a morality ‘compatible with very great iniquity’ is a morality in name only).
Following Hart, Kramer grants the utility of Fuller's criteria as threshold conditions for the existence of the rule of law, but argues that once this threshold is met the question of closer compliance to an ideal legal system does not apply.
Above that [threshold] level, up to some considerably higher point, any further compliance with each principle will enhance the clarity and robustness of the status of a legal system as such but will not be indispensable for the very applicability of that status.Footnote 31
If legality is a threshold property, as Kramer argues, it is not a matter of degree but a quality which is either absent or present. Defects of universality, knowability, followability and justifiability will only reflect on the overall legal order if the lacunae they introduce are so extensive that the rule of law ceases to obtain; this in turn will only be the case at the point where the normalisation of exceptional measures has proceeded so far that laws in general are particularistic, unknowable, unfollowable and/or unchallengeable.Footnote 32 Short of that dystopian point, the worst that these defects can do is reduce the clarity with which the rule of law is identifiable as operating. This lack of clarity may in turn be justified by an appeal to other values, even non-legal values such as security or efficiency.
The effect of the ‘threshold’ model is thus to disable the ‘counter-law’ critique by dividing it in two. At the point where deviations from Fuller's criteria are so great as to impair the operation of the legal system, a technical critique can be mounted, inasmuch as sufficiently gross defects will prevent the legal system from functioning at all. Where defects do not rise to that level, critiques based on Fuller's criteria are based on a freestanding morality and have no grounding in the nature of law.
Conversely, Simmonds presents the model and its relation to reality in terms of an archetype:
To count as an instance of law, a regime must approximate to the archetype to some degree: it must, so to speak, participate in the form of law. Yet the very fact that such participation can be instantiated to varying degrees means that the archetype can nevertheless constitute a guiding ideal to which legal systems ought to strive to conform more closely.Footnote 33
Fuller's criteria should be understood as sketching out the key features of an approach to governing human societies – an ideal to which real systems of law may approximate more or less closely. Notably, Simmonds also suggests that ‘the possibility of revision is inherent in any statement of a guiding archetype’:Footnote 34 to understand the rule of law as an archetype brings with it the necessity of a process of reflective revision, and the recognition that any institutional implementation of the rule of law is likely to be imperfect. The archetype offers guidance for the development of the law; the law in this sense is ‘the process of its own becoming’.Footnote 35 Although never fully achieved, the rule of law represents an end-point towards which progress can always be made.Footnote 36
The archetypal model, Simmonds argues, centres on ‘the idea of a domain of universality and necessity in human affairs, providing a degree of independence from the power of others’.Footnote 37 To say that the rule of law exists is to say that ‘domains of liberty that are independent of the will of anyone’Footnote 38 have been secured, more or less effectively, by law.Footnote 39 Hence, the more fully the rule of law is realised – which is partly, but not exclusively, a matter of conformity to Fuller's criteria – the more fully the law guarantees individual freedom. Autonomy is the core value, to be maximised by the operation of laws approximating to Fuller's criteria; indeed, there is a mutually reinforcing relationship between law and (a certain form of) autonomy.Footnote 40
The ‘threshold’ and ‘archetype’ models can also be distinguished by their response to the relativising critiques discussed above. Lacey notes that the rule of law was recognised as an ideal long before it was feasible to realise Fuller's requirements of publicity and congruence with official action.Footnote 41 The threshold model would concede this point, but at the cost of challenging historical actors’ self-understanding, denying the name of ‘rule of law’ to formulations of that concept which did not meet Fulleran criteria. Lacey argues for some continuity over time, but only at the level of ‘a modest and formal conception of the rule of law’, denoting social arrangements ‘with the dual capacity to constrain or temper and hence legitimize power’.Footnote 42 The ‘archetype’ model offers the possibility of a fuller picture of the normative development of the rule of law, situating changing conceptions within an overarching conceptual framework without according any privilege to the present day; a world in which individuals appear before the law as autonomous moral agents ‘has arguably never existed’Footnote 43 in fully realised form, any more than (for example) a fully realised democracy. The rule of law in archetypal perspective is thus ‘the enterprise of subjecting human conduct to the governance of rules’,Footnote 44 considered as a long-term project and one very far from adequate realisation. This model presents the rule of law not only as a normative standard but as a socially-recognised ideal, and as such lends itself to the classification both of existing legal orders and of movements towards or away from the rule of law.
The archetypal model also enables us to understand counter-law in terms of the selective suspension or reversal of that project. To the extent that counter-law is an active tendency, we might then expect to see four counter-trends to the realisation of the rule of law. Against universality, counter-law offers selectivity: selective criminalisation systematically subjects some groups more effectively than others to the coercive powers of the criminal justice system. Against knowability, counter-law offers laws designed to be targeted retrospectively: laws that can be used to criminalise but cannot be used by individuals to guide their behaviour. Against followability, counter-law limits the domain of free choice, either by directly controlling behaviour or by loading individuals with positive obligations. Against justifiability, counter-law offers laws drafted so broadly – and, on occasion, procedures controlled so tightly – as to be beyond the scope of effective appeal, taking these offences outside the circuit of challenge, review and revision. The effect of these conjoined tendencies is to attack the values of individual autonomy and moral agency, annulling the ‘domains of liberty’ within which individuals should be able to pursue their own life plans.
Situation, preparation, inchoation: problematic trends in counter-terrorism
The new criminal offences introduced by the Terrorism Acts of 2000 and 2006 exemplify the development of counter-law. Three types of offence are discussed below: inchoate offences, preparatory offences and situational offences; all of these fall into the larger family of preventive offences identified in Ashworth and Zedner's typology.Footnote 45 These offences all tend to dissociate counter-terrorism from criminal law's standard requirement for ‘a clear illegal act that is committed with fault’,Footnote 46 focusing instead on a pattern of behaviour identified as undesirable. The importance of the offences under discussion in contemporary counter-terrorist criminal justice is hard to overstate. The Crown Prosecution Service website lists 155 successful counter-terrorist successful prosecutions between the passage of the Terrorism Act 2006 and the end of calendar year 2016, involving the bringing of 391 charges of which 345 were brought to conviction.Footnote 47 The specifically counter-terrorist charges brought in these cases – all of which can be classified as inchoate, preparatory or situational – account for 66% of all charges brought; much of the remainder is accounted for by general inchoate and situational offences (eg conspiracy and possession offences, respectively).
Inchoate offences criminalise acts oriented to achieving a wrongful end, where those acts are incomplete or unsuccessful. The term is here being used to cover both general inchoate offences (attempt, conspiracy and encouragement or assistance) and statutory offences defined in the inchoate mode.Footnote 48 The inchoate offence type which has seen the most counter-terrorist legislative innovation is encouragement. Under the Terrorism Act 2006 it is an offence to publish a statement which ‘glorifies’ the commission or preparation of acts of terrorism (in the past, in future or generally), in such a way as to portray those acts as worthy of emulating, if the person doing so either intends that the audience should be encouraged to commit, prepare or instigate acts of terrorism as a result or is reckless as to this possibility (Terrorism Act 2006, s 1). It is also an offence to distribute or circulate any publication which is likely either to encourage carrying out acts of terrorism or to offer practical assistance in doing so; again, this is the case whether the publications are circulated intentionally or merely recklessly, with respect to their likely effect (Terrorism Act 2006, s 2).
Three points should be noted here. First, in the case of the s 1 ‘glorification’ offence the encouragement may be indirect and/or incomplete: thanks to the ‘instigate’ clause, those encouraged need not be encouraged to commit terrorist acts themselves, even with the Terrorism Act 2000's expansive definition of terrorist acts to include threats. If A makes a ‘terrorist’ threat after being encouraged to do so by B, who had previously read material in praise of terrorist acts circulated by C, C may be found guilty of a terrorist offence (with a maximum sentence of seven years).Footnote 49 Secondly, in criminalising reckless conduct, both the ‘glorification’ offence and the ‘encouragement’ arm of the s 2 publication offence go beyond more usual definitions of encouragement to commit an offence.Footnote 50 Although the recklessness required by the s 1 and s 2 offences is subjectiveFootnote 51 – so that the offence requires knowingly disregarding a substantial risk of encouraging terrorist activity – the burden remains on the defence to demonstrate that the statement made or material circulated did not have her endorsement, a non-endorsement which in practical terms is likely to require a positive action (ie the furnishing of appropriate disclaimers at the time a document is circulated).
Thirdly, not only do these offences not require that any terrorist action (or instigation) takes place as a result of the publication or circulation; they do not require that ‘any person … is in fact encouraged’ (Terrorism Act 2006, s 2(8)). The publication and circulation offences are doubly inchoate: not only is there no need for any individual to be successfully ‘encouraged’ to the point of attempting a terrorist act; there is no need for the publication or circulation to succeed in producing the state of ‘encouragement’ in any individual. The conceptual horizon of terrorist offences thus extends back in time and down the scale of severity, identifying ‘terrorism’ as a culpable pattern of activity rather than in terms of any readily identifiable harm done or prohibition breached. In terms of severity, the definition explicitly includes acts which were not only harmless but had no wrongful intent. In temporal terms, the offence is the publication rather than the possible subsequent encouragement, let alone any action which might have been encouraged. The logic seems to be that a pattern of behaviour identifiable as ‘terrorism’ is afoot, which should be interrupted by criminalising anything associated with it.
Preparatory offences, secondly, go even further than inchoate offences in extending the scope of criminal liability back in time and down the scale of harm. The term is being used here to cover statutory offences which consist of undertaking activity – which may in itself be both harmless and legal – in preparation for a distinct future criminal act.Footnote 52 Thus the Terrorism Act 2000 criminalises training in weapons and explosives – provisions expanded under the Terrorism Act 2006 to include training in ‘the design or adaptation for the purposes of terrorism … of any method or technique for doing anything’ (Terrorism Act 2006, s 6(3)(c)), a formulation which at least has the merit of leaving little work for future legislators. The most frequently used preparatory offence in this area – and the one which most vividly illustrates the extension of the scope of counter-terrorist legislation – is ‘Preparation of terrorist acts’, also introduced in the 2006 Act (Terrorism Act 2006, s 5). If an individual has the intention of either committing or assisting a terrorist act, ‘any conduct in preparation for giving effect to his intention’ can constitute an offence, potentially carrying a life sentence. The ‘preparation’ offence thus makes it possible to criminalise individuals who have not committed (or, in the normal sense of the word, attempted) anything recognisable as a terrorist act, on the basis that an identifiable, individually innocuous action formed part of a pattern of behaviour which might in future have led to the commission of acts of terrorism. This can even be done when that pattern of behaviour has been interrupted by police intervention, potentially removing any basis for distinguishing between behaviours which would and would not ultimately have led to harmful acts – other than by accepting the interpretation advanced by prosecutors.
Problems with preparatory offences are legion. Crucially, they risk criminalising actions on the grounds of their potential outcomes – outcomes which may depend on a further intervening decision, if only the decision to persist in a chosen course of action. In Ashworth and Zedner's words, ‘to hold a person liable now for her possible future actions … is objectionable in principle’.Footnote 53 Ramsay argues further that such offences ‘discount the formal agency or personhood of offenders’Footnote 54 by eliminating any time for reflection, criminalising on the basis of ‘a will to commit dangerous acts’Footnote 55 and effectively working on the basis that ‘[i]t is not the act but the actor that is dangerous’.Footnote 56 Hallevy's extreme and draconian proposal to cut the ‘preparation’ knot by bringing preparatory activities under the heading of ‘attempt’ merits notice here. For Hallevy an attempt is formed when the suspect first decides to commit an offence; once that decision has been acted on in any way ‘the attempted offense has been constituted’.Footnote 57 The putative offender can – and should – be prosecuted on the basis of the dangerousness he has already manifested: he ‘became a danger to society when he made his decision’.Footnote 58 The breadth and flexibility of the s5 preparation offence, in particular, evokes Hallevy's logic: once a suspect is identified as harbouring dangerous intentions, any conduct in preparation for putting them into effect can be criminalized, even if investigation of this conduct does not yield ‘any blueprint, attack plan or endgame’Footnote 59 (in the words of an officer involved in investigating the Farooqi groupFootnote 60). The offence facilitates prosecution of conduct that had not risen to the level of culpability, and as such falls into Edwards's category of ‘ouster’ offences:Footnote 61 offences defined so as to be provable against a much wider range of people than the offenders actually being targeted. The effect of such offences is to redefine the role of the courts: the courts are ousted from their role of determining guilt by being presented with offences defined in such a way that only one verdict is possible, while the actual wrong for which the accused is being prosecuted goes untested.
Situational offences, finally, are a group of offence types which criminalise a state of affairs rather than an action: as well as offences where the actus reus is being present at a certain location, the term is being used here to include membership offences, possession offences and offences of omission, eg failure to inform. Situational offences, in effect, replace the commission of a wrongful act with the existence of a wrongful state of affairs – although, as Simester argues, the offence can be viewed in terms of an implicit wrongful act which was the precursor to that state of affairs (‘[i]t is enough that there be some actus reus event over which the agent has control, that it is not impossible for the agent to prevent’).Footnote 62 The absence of any requirement to prove a positive action greatly strengthens the position of the prosecution. The factual element of the offence can be attested by the prosecution: if the arresting officer confirms that the accused was in fact in possession of an incriminating item (etc), this evidence will not generally be subjected to any credible challenge. The mental element of a situational offence is typically limited to a defence, eg that the accused was ignorant of the item's presence or had a lawful purpose in taking possession of it. This tends to reverse the onus of proof, calling on the defence to bring evidence in support of the innocence of the accused rather than demanding that the prosecution prove guilt. In some cases the mental element of a situational offence may even be assimilated to the factual element. The offence of ‘Attendance at a place used for terrorist training’ introduced by the Terrorism Act 2006 (maximum sentence ten years’ imprisonment) is taken as proven if a person attending the relevant place at the relevant time ‘could not reasonably have failed to understand’ that training connected with terrorism was taking place (Terrorism Act 2006, s 8); given the right combination of circumstances, the mental element of the offence can simply be assumed.
A partial reversal of the burden of proof, in situational offences, may be acceptable when the undesirable behaviour at which the law aims is, precisely, the existence of a particular state of affairs – eg the possession of illegal weaponry or membership of proscribed organisations. Whether simple possession of any article, however dangerous, should in principle be criminalised can be debated.Footnote 63 If that point is conceded, however, the mental element may in some situations be assumed,Footnote 64 or addressed by placing an evidential burden on the defence.Footnote 65 Reversing the burden of proof is more problematic where the state of affairs being addressed is innocuous in itself, but is criminalised as a precursor to other undesirable behaviour. While walking down the road carrying a toolbox is a harmless and lawful activity, carrying tools for use in a burglary in a public place constitutes the offence of ‘going equipped for stealing’ under the Theft Act 1968 – and proof that the accused had the item in his or her possession is admissible as evidence that the item was intended for use in burglary or theft, placing an evidential burden on the defence (Theft Act 1968, s 25). The scope of ‘non-constitutive’ offences such as these may reasonably cause concern.Footnote 66
The main counter-terrorist situational offences are possession offences, specifically the offences created by sections 57 and 58 of the Terrorism Act 2000: possession of an article (any article) ‘in circumstances which give rise to a reasonable suspicion that his possession is for a purpose connected with the commission, preparation or instigation of an act of terrorism’; and possession of a document or record ‘likely to be useful to a person committing or preparing an act of terrorism’. Needless to say, the wording of these offences is extraordinarily broad. The scope of the s 57 offence has been reduced slightly by the ruling in R v Zafar and Others.Footnote 67 that the ‘a purpose connected with’ wording should be read as if it required a direct connection between the object possessed and the commission, preparation or instigation of terrorism. Even with this qualification, the offence is effectively one of arousing suspicion, giving the defence the burden of giving an account that will allay those suspicions. The burden of proof similarly lies with the defence in the case of s 58 offences: it is up to the defence to allay suspicion by proving the existence of a reasonable excuse.Footnote 68 Moreover, in R v G and J Footnote 69 the House of Lords ruled that any such excuse must refer to a course of conduct that was in itself reasonable: a burglar, apprehended holding plans of the local police station and wrongly accused of terrorist activity, can be found guilty under s 58 regardless unless he can demonstrate that he had lawful reasons for holding that document.
A situational offence is also at issue where a defendant can be charged with having remained, guiltily, silent, as in the case of the offence of failing to disclose to the police ‘as soon as reasonably practicable’ any information which might prevent an act of terrorism or lead to the arrest of a terrorist (Terrorism Act 2000, s 38B (inserted by Anti-Terrorism, Crime and Security Act 2001, s 117)). As with other situational offences, it will not be hard to satisfy a jury as to the factual element of the offence, given a sworn statement that the individual did in fact have the information; the mental element (ie the lack of a reasonable excuse for remaining silence) can also be assumed, subject to disproof by the defence. Offences such as this push Simester's ‘event over which the agent has control’ formulation to its limit, effectively making an actus reus out of the agent's failure to resolve a culpable situation. Although the construction of non-disclosure as an offence can be defended,Footnote 70 the overriding impression is not of the criminalisation of a wrongful act, but of the use of criminalisation as a tool to interrupt a loosely-defined pattern of behaviour.
What is against this law? What is this law against?
Two closely-related themes recurred through the previous section: the refusal of definition and precision in defining identifiably wrongful conduct, in favour of legislation targeting loosely-defined and temporally extended patterns of behaviour; and the utility of legislation designed on these lines for criminalization and prosecution purposes. Perhaps surprisingly, the impression of vagueness by design is not dispelled by s 1 of the Terrorism Act 2000, in which ‘terrorism’ is defined with every appearance of precision. The definition of terrorism in the 2000 Act replaced the definition brought into law by the Prevention of Terrorism (Temporary Powers) Act 1974; this had defined terrorism laconically as ‘the use of violence for political ends, [including] any use of violence for the purpose of putting the public or any section of the public in fear’.Footnote 71 The much lengthier 2000 definition has three elements, which itemise the acts which can be identified as terrorism; state the political or other ideological motive which must underlie them; and specify the tactical purpose with which they are committed.Footnote 72 This third element (specifying the goal of influencing a government organisation or intimidating the public) does not need to be proved if the act included the use of firearms or explosives (Terrorism Act 2000, s 1(3)).Footnote 73
Each element of the definition is problematic. Whether ideological motivation is necessarily an aggravating factor – elevating a crime of violence, or a mere threat of violence, to the rank of terrorist offencesFootnote 74 – is debatable;Footnote 75 nor is it clear that either intimidating the public or influencing the government has ever been central to terrorist activity.Footnote 76 As for the ‘act’ definitions, it is worth noting that the formulation ‘the use or threat’ precedes and applies to each of them – a formulation apparently derived from a definition used by the FBI.Footnote 77 Although not widely commented on at the time,Footnote 78 this was a major innovation. While the ‘use or threat’ formulation is not unfamiliar in English criminal law,Footnote 79 the range of the activities specified in the 2000 Act made it particularly powerful here: threatening to endanger life or threatening to create a risk to health and safety (among much else) could henceforth be classified as an act of terrorism in its own right. This is a striking extension of the concept of terrorism, both in temporal scope and in the range of activities covered by it. If the threat of an action is just as much an act of terrorism as the action itself, the criminalisation of the act can clearly not be justified by harm done or even by an appeal to the wrongfulness of the act in itself, unless on the basis that a threat of harm is inherently wrong.Footnote 80
At issue here are two separate shifts in the definition of terrorism, as compared with the 1974 definition. Detonating a bomb that causes £700 million of property damage (as in Manchester in 1996) or kills 22 people (as in Manchester in 2017) is unquestionably a harmful and wrongful act, and as such can be caught by the 1974 Act's reference to ‘violence’. The 1974 definition effectively starts from a group of actions which would have constituted criminal offences in any case, by virtue of their harmful effects, then demarcates a politically-motivated subset of those offences as especially blameworthy acts of ‘terrorism’.Footnote 81 By contrast, the 2000 Act's definition tends to define terrorism primarily in terms of wrongful rather than harmful conduct, and in terms of patterns of behaviour rather than discrete acts.
These offences share another, curious attribute: they all depend on the initial ascription of terrorism to the accused. The clauses in the 2000 Act which purport to define terrorism allow for counter-terrorist powers to be invoked both broadly and selectively. Among the counter-terrorist prosecutions documented by the CPS is the case of Ryan McGee, a vocal racistFootnote 82 who built a viable bomb. McGee was prosecuted under the Terrorism Act, but only under s 58 (possession of information potentially useful to terrorists); his bomb-making activities were dealt with by a charge under the Explosive Substances Act 1883. He was sentenced to two years’ imprisonment for the explosives charge, with a 12-month sentence for the s 58 charge, to run concurrently.Footnote 83 Terrorism – as defined in the 2000 Act – encompasses the making of politically-motivated threats of serious violence involving explosives. What qualifies as an action in preparation for such an act is undefined in law, but amateur bomb-making carried out by a political extremist would seem to qualify. Presumably the reason that McGee's actions were not constructed as terrorism is simply that they had not initially been viewed as ‘terrorism’, or the suspect as a ‘terrorist’: according to news reports, the Crown Prosecution Service declined to bring ‘preparation’ charges on the grounds that ‘it was never McGee's intention to use the device for any terrorist or violent purpose’.Footnote 84 A suspect's intentions must be reconstructed from a range of evidence, including patterns of behaviour apparently directed towards realising them as well as – or to the exclusion of – first-person testimony; compare R v Tabbakh,Footnote 85 in which the defendant's disavowal of any terrorist purposes (in assembling a rudimentary and non-viable device) was disregarded.Footnote 86 Greene notes the opacity of subjective intention: ‘the label of “terrorist” may facilitate the inference of intention’.Footnote 87 The risk is that, rather than labelling following the classification of an offence according to its definition, the classification of the offence – and the charges brought – are determined by ascriptive labelling of the suspect.
Terrorism in the law is effectively defined as a pattern of behaviour which it is vitally important to disrupt; which can only reliably be identified by specialists in counter-terrorism; and which can be identified by those experts in the absence of any identifiable wrongful act or any proven guilty intent. The effects of this built-in reliance on administrative discretion are potentially grave: ‘in practical terms the outcome is no different than if terrorism is not defined at all’.Footnote 88 The logic of the additional offences created since 2000, with their focus on the inchoate, preparatory and situational, follows from the original definition: these are not laws that criminalise identifiable wrongful acts, but laws that make it possible to use criminalisation as a tool for selectively pre-empting potential future acts.
Law, counter-terrorism, counter-law
These peculiarities of counter-terrorist legislation map surprisingly well onto the archetypal model of law developed above, albeit in negative. Knowability is a key feature of law as a guide to behaviour. Reliance on prosecutorial discretion in such a fundamental matter as the application of the label of ‘terrorist’ is unsatisfactory, particularly when the labelling is retrospective. Both preparatory and situational offences also fall short in this respect. A law prohibiting the possession of articles which might subsequently be presented as useful for the commission of an act of violence, or the commission of acts which might subsequently be presented as having been preparatory to what might have been a future act of violence, is not a law whose scope can be reliably known. The rule of law demands that the scope of terrorist offences should be knowable in advance.Footnote 89
Followability is also problematic, and not only to the extent that it is dependent on knowability. The ‘glorification’ and ‘encouragement’ offences effectively prohibit communicative acts which might later be presented as having had a tendency to encourage potential future acts of violence, even if no violence or even encouragement had taken place. Someone who may at some future date be charged with breaking these laws may choose to regulate her behaviour so as to reliably avoid it, but only by conforming her behaviour to the precise requirements of the offence. To avoid the charge of glorifying terrorism, in other words, does not require choosing not to glorify terrorism; rather, it requires the affirmative act of a clear and detailed disclaimer. A law that hedges lawful and harmless conduct about with additional requirements violates the spirit of followability. We may recall that the Control Order regime was also widely criticised – and individual control orders overturned – for placing suspects under conditions which impinged on their liberty to the point of not being practically followable, in the sense of allowing a free choice to obey:Footnote 90 again, the identification of an individual as a terrorist suspect placed her under specific positive obligations in respect of a broad area of life.
Of these four demands, justifiability is perhaps the one which contemporary counter-terrorist legislation meets most effectively; appellate rulings have clarified (and restricted) the scope of the Terrorism Act 2000, s 57 offence (R v Zafar and Others Footnote 91) and the recklessness element of the Terrorism Act 2006, s 2 offence (R v Faraz Footnote 92), as well as querying the definition of terrorism itself (R v Gul Footnote 93). (However, clarification has effectively extended the reach of the Terrorism Act 2000, s 57 offence (R v G and J Footnote 94).) That said, the possibility of justification is limited when provisions are drafted so as to leave no room for challenge, as in the Terrorism Act 2006, s 8 ‘training’ offences. Further, justification is challenged directly when trials are held wholly or partly in camera,Footnote 95 or when individuals are not allowed to hear the evidence against them, as in the proceedings of the Special Immigration Appeals Commission.Footnote 96
Contemporary counter-terrorist legislation also breaks with the most fundamental requirement of Fuller's morality of law, and of thinner concepts of the rule of law:Footnote 97 the requirement of universality. Failure to specify the conduct targeted fully necessitates reliance on prosecutorial discretion: some individuals are prosecuted for conduct answering to the offence descriptions and others not. The point generalises: some individuals live with the awareness that they are liable to be prosecuted for conduct which may appear to answer to the offence descriptions; others – a majority – can be confident that they will not. In effect, the law targets an identifiable minority, treating individuals within it ‘as objects of control rather than as citizens to be engaged with’.Footnote 98 Membership offences are particularly questionable under this heading: to proscribe an organisation associated with a particular community must necessarily impose a disproportionate burden, and risk of incrimination, on members of that community as compared to the general public, particularly when the decision to proscribe may be poorly grounded.Footnote 99
The rule of law gives citizens ‘domains of optional conduct’, secured with a ‘protective perimeter’ of individual rights:Footnote 100 I can do what I like as long as I do no wrong to anyone else, and what counts as doing wrong is defined by the rights which each person is deemed to have (bodily integrity, private property, freedom of speech etc). Once an individual has been identified as a potential terrorist, however, a vast range of behaviour comes under suspicion. Moreover, to the extent that labelling practices are predictable and patterned, individuals belonging to those population groups which are disproportionately likely to be labelled will also find their activities affected, if only through the chilling effect of internalised suspicion. It is meaningless to say that the law protects freedom of expression or private property – let alone that it protects individual freedom of action – when it selectively penalises particular examples of free expression, property ownership and harmless behaviour. The legal regime under which a terrorist suspect – or someone who believes that she is liable to be viewed as a terrorist suspect – lives is one with no protected domain of optional conduct and no reliable rights. Counter-terrorist counter-law annuls the rule of law.
Four counter-arguments to this conclusion suggest themselves. First, prosecutorial discretion may be invoked: artificial examples of potentially criminalisable actions – eating healthilyFootnote 101 or wearing clean clothesFootnote 102 considered as ‘conduct in preparation for giving effect’ to terrorist intentions – are and remain absurd, it is argued, because no prosecution of such conduct would ever get to court. Conversely (secondly), it may be argued that the guidance given by the law to individuals intent on terrorism is both knowable and followable: anyone preparing terrorist acts should stop. Moreover, even if the guidance offered to those individuals were overbroad, targeting pre-preparatory as well as truly preparatory actions, this would not deprive society of any useful or constructive activity and hence would not be regrettable. Third, the case for prosecutorial necessity may be urged: this is the argument that the activities targeted by counter-terrorist legislation, if allowed to proceed, pose so grave a risk to society as to necessitate the use of the criminal law with pre-emptive effect. (This and the first argument both in different ways evoke Macdonald's ‘investigative efficiency’Footnote 103 ideal-type, which stresses the reliability of police investigators and their findings.) Lastly, it can be argued that the rule of law maintains an adequate – threshold – level of operation, even for individuals who may come under suspicion of terrorism; this position echoes Macdonald's ‘adversarial reliability’ ideal-type,Footnote 104 which stresses the reliability of the courts in processing fallible investigative findings.
The ‘prosecutorial discretion’ argument reminds us, correctly, that criminal law ‘in the books’ is only one element of the process of criminalisation and should not be studied in isolation.Footnote 105 However, the trust which we are invited to place in prosecutorial discretion is, by its nature, hard to validate. If we had perfect knowledge that unreliable and tenuous examples of ‘conduct in preparation’ would never get to court, we could also have perfect confidence that those examples that do get to court were reliable. In the absence of that knowledge – in the absence, among other things, of full data on unsuccessful and abandoned prosecutions as well as successes – we are being invited to take both propositions on trust. There is also the danger that normalising prosecutorial discretion ‘permits overcriminalization, which in turn encourages more discretion’, resulting in ‘an unwritten criminal “law” that consists only of enforcers’ discretionary decisions’.Footnote 106 The remedy entrenches the problem: prosecutorial discretion may mitigate the worst effects of broad and vague legislative wording, but reliance on discretion takes the law in practice further away from the law in the books, and hence militates against knowability and followability.
According to the ‘adequate guidance’ argument, even if the actions of potential terrorists are restricted more than is entirely necessary, this loss is by definition of no social value.Footnote 107 The problem with this argument is that it conflates the subjective (but unknowable) reality of an intentional state with intention imputed on the basis of observed statements and activities. In other words, once an individual has been identified – rightly or wrongly – as harbouring terrorist intentions, the breadth of the statutory offences throws suspicion on a wide range of that individual's acts, particularly acts intended to advance a ‘political, religious, racial or ideological cause’ (in the words of the 2000 Act). Moreover, a similar effect operates for putatively suspicious individuals themselves: the effect of the legislation is to create a chilling effect, tending to discourage and delegitimise activities, and the expression of views, that might (at a later date) be regarded as (having been) suspicious. This is the case both for those individuals who have fallen under suspicion for good reason and for the much larger group who fear being wrongly or falsely accused.
The ‘prosecutorial necessity’ argument again recalls us to the law in practice. If McGee was not charged with a preparation offence where TabbakhFootnote 108 was – receiving a sentence of seven years to McGee's two – this does not mean that the law looks more kindly on a white racist who succeeds in makes a bomb than on an Islamist who fails to do so; it simply means that prosecutors, with the information and intelligence available to them, took the view that McGee's bomb-making was not part of a pattern of behaviour that would eventually have led to terrorism, while Tabbakh's attempted bomb-making was. Similarly, if a suspect is found guilty of preparation for terrorism on the basis that he had had incriminating conversations with undercover police officers,Footnote 109 or that he had expressed willingness to use his position of employment for terrorist purposes,Footnote 110 this does not mean that people are being convicted of terrorist offences on the basis of little more than loose talk. On the contrary, it can be argued, what this brings home to us is the necessity of a law drafted flexibly enough to make it possible to convict dangerous individuals such as Tabbakh, Farooqi and Karim, even when they have not furnished prosecutors with any discrete and definite actus reus.
The problem with this argument is that, even if we assume that its factual basis is correct, the original critique is not touched by it. Any legislation worded loosely enough to be used in this selective and instrumental way will, by the same token, be worded too loosely to be either fully knowable or effectively followable, by those individuals who fall (or are seen as likely to fall, or believe themselves likely to fall) under suspicion of terrorist activities or sympathies. As for whether such a deviation from the rule of law might be an acceptable price to pay for increased security from terrorism, the point is to name the price: in this case, the awareness that the law is less effectively universal (hence more discriminatory) than it had been – and, as it applies to a minority of people, less effectively knowable, followable and justifiable. Trading the security of the majority against the rights of the minority in this way surely cannot be acceptable.Footnote 111
According to the ‘adequate threshold’ argument, lastly, terrorist suspects do not live outside the rule of law: it and its associated procedures and guarantees offer their protections to all who need them, terrorist suspects included. There is, still, no punishment without a crime to justify it and no crime without a law to define it; procedural justice extends to terrorist suspects on trial, and if their conduct is genuinely unproblematic they will not be convicted. As for the potential terrorist suspect who has not been – and may never be – in contact with the criminal justice system, she knows (as a citizen of a law-bound society) that the law leaves her free to act in any way he or she pleases, as long as other people's rights are not infringed. However, she also knows that it would be advisable to avoid doing anything, possessing any item, attending any location, acquiring any information, advocating any action, making any statement, failing to disown any statement by a third party or failing to pass any information to the police where this action, item, attendance, information, statement or non-statement might be presented as being preparatory to, supportive of or otherwise related to terrorism. For anyone who is (or believes herself to be) liable to fall under suspicion of involvement in terrorism, counter-terrorist prohibitions overshadow a substantial area of social and political life. (This consideration also makes the guarantees available to the terrorist suspect within the system considerably less substantial.) Crucially, the precise extent and boundaries of the area of life within which behaviour may potentially be found suspicious are undefined, meaning that in effect its scope is unlimited.
Conclusion
Counter-terrorist legislation since 2000 has been designed for the selective criminalisation of individuals deemed to be engaging in patterns of behaviour labelled as related to terrorism, so as to interrupt those patterns of behaviour and pre-empt the commission of anything rising to the level of a terrorist act. By the same token, it has developed as a body of law that effectively applies to a minority but not to the majority; that criminalises behaviour, by those who are of interest to counter-terrorist investigators, in terms so broad and so loose as to give minimal guidance to those individuals; that frequently requires suspects to take positive action to demonstrate conformity with the law; and that is drafted to offer minimal handholds for appeal and challenge. In short, the law deviates – by design and not merely in practice – from the rule of law's values of universality, knowability, followability and justifiability.
What is to be done? The foregoing critique is consistent with – and hopefully adds normative weight to – Zedner's critique of the ‘terrorising’ of criminal justice and her proposed tenets for dealing with terrorist offences within the ‘criminal justice model’.Footnote 112 We might add that the criminal law should, in Duff's terms, deal honestly with those subject to it, defining as crimes only ‘public wrongs’Footnote 113 that each person is in principle capable of recognising and free to avoid. Any future reform of counter-terrorist legislation should thus start from the principle of creating offences which preserve the equal freedom of the rationally and morally autonomous individuals subject to them, and which therefore adhere as far as possible to the four rule-of-law principles discussed above. This is in part a call for parsimony. The 1974 Prevention of Terrorism (Temporary Powers) Act, for all its faults, created only a handful of new offences (connected with membership or support of proscribed organisations) – and this element of the Act was described by the then Home Secretary as ‘for show’.Footnote 114
Deviation from rule of law values will sometimes be necessary, just as it is sometimes necessary to restrain a dangerous individual who has not yet committed a culpable act.Footnote 115 What is crucial is that any such deviation is openly acknowledged and justified; the pre-2000 model of temporary legislation in response to a specific emergency – now discredited through over-extension – may be due for a revival.Footnote 116 Above all, there should be more awareness of the distinctness – to put it no more strongly – of the roles of policing and the criminal law;Footnote 117 legislation should not be used to give a legal armature to counter-terrorist policing.
Lastly, this paper is offered as a contribution to two related projects. On one hand, the elaboration of concepts of legality and the rule of law can benefit from a demonstration of how they can be deployed in contemporary policy debates, providing normative resources for the critique of illiberal and authoritarian developments. On the other, the historicist, dialectical and pluralist critiques of the liberal legal model can ultimately be strengthened by acknowledgment of the power and relevance of the norms they critique. Let liberal ideals of legality be subjected to critical scrutiny, returning them to their historical context and drawing out their blind spots and contradictions – but let them be attacked at their strongest points as well as their weakest, engaging with the critical and emancipatory power that those ideals have. Let us look not merely to the negation of the liberal model but to the preservation of its positive content, in and through its ultimate formal supersession.
