Introduction
In the diverse New Institutionalism literature, the focus of analysis in explaining complex political change is institutions, which are viewed as structures that can limit or shape policy options in a given context. As I will show below, it has become accepted that norms can act in institution-like ways, by creating “elements of order and predictability”, and by enabling and constraining “actors as they act within a logic of appropriate action” (March and Olsen Reference March, Olsen, Binder, Rhodes and Rockman2008, 4). A norm that has, to date, not been examined in terms of its capacity to act in institution-like ways is freedom of speech. Freedom of speech is a highly valued norm that is often – indeed routinely – affirmed by policymakers and political leaders as one of the values of liberal democratic societies. So, can freedom of speech be conceptualised as acting in institution-like ways? Relatedly, what does conceptualising free speech as an institution illuminate about processes of policy change?
In order to answer these questions, first I will clarify how I conceptualise free speech as a norm and how it might be understood as an institution. In doing so, I will draw out two definitions of norms acting in institution-like ways from the literature: a broad and a narrow definition. I then outline and justify my case selection and the research design. In the next section, I trace and analyse policy change between 2001 and 2011 in the United States (US), the United Kingdom (UK) and Australia, a period during which significant new limits on freedom of speech in relation to speech regarded as harmful to national security were introduced. This account will show how free speech acted in institution-like ways to constrain, shape and enable policy change, and how it interacted with and was influenced by other institutions. In the following section, I show how this analysis exposes the complexities of policy change by revealing three things. First, it reveals that norms that act in institution-like ways can simultaneously act to constrain, shape and enable change, and be subject to change in interaction with other institutions. Second, it demonstrates how a broad, cultural level institution can mask policy change at the narrow, rule-based level even where the latter contradicts and is incompatible with the former. Finally, the analysis shows that complexity and variation in speech regulation can be understood as consequences of the to-be-expected variability in the institutionalisation of a norm.
Conceptualising free speech
Free speech can be understood as a norm, where a norm is defined as an idea that reflects “fundamental values, organizing principles or standard procedures that resonate across many state and global actors, having gained support in multiple forums including official policies, laws, treaties or agreements” (Krook and True Reference Krook and True2010, 103–104). The norm of free speech is a widely supported value in democratic societies, which has been instantiated in constitutions, statute and common law, and is recognised in Article 19 of the Universal Declaration of Human Rights and Article 19 of the International Covenant on Civil and Political Rights.
Can free speech also be conceived of as an institution? One of the characteristics of the New Institutionalist literature as it has developed since the 1980s is its acceptance that institutions are to be conceived broadly as structures of expectations and traditions that are “relatively resilient to the idiosyncratic preferences and expectations of individuals” (March and Olsen Reference March and Olsen1984, 741), and that constrain and shape policy change. The concept of an institution is inclusive of informal institutions (Mackay, Kenny and Chappell Reference Mackay, Kenny and Chappell2010, 576) and ideational structures (Bell Reference Bell2012, 666).
In spite of the different approaches, analytical foci and ontologies of varieties of New Institutionalism, institutions have come to be defined in ways that recognise that, under some circumstances, norms can act in institution-like ways (e.g. Goodin Reference Goodin and Goodin1996, 19; Streeck and Thelen Reference Streeck, Thelen, Streeck and Thelen2005, 9). This is not to suggest that all norms can be considered to be institutions, since to do so would elide the differences between them. Nevertheless, Streeck and Thelen have argued that although different theorists will have broader or more narrow understandings of institutions, they all understand institutions as including those norms and rules in relation to which, “there are strong enough sanctions against deviating from them” (Reference Streeck, Thelen, Streeck and Thelen2005, 10).
But differentiation persists in how some norms can be considered to be institutional in character, with historical institutionalists tending to define institutions as inclusive of norms that have formal status (Hall Reference Hall1986, 19) or are structurally routinised, meaning that they are “embedded in the organizational structure of the polity” (Hall and Taylor Reference Hall and Taylor1996, 938). Other historical institutionalists have defined institutions as inclusive of “social norms” (March and Olsen Reference March and Olsen1989, 52, 22), or as “formal rules, policy structures or norms” (Pierson, cited in Thelen Reference Thelen1999, 382). This recognises that norms can act in institution-like ways when they constitute formal and informal rules and sanctions.
By contrast, sociological institutionalists tend to utilise a broader concept of institutions that includes “cultural conventions, norms and cognitive frames” (Hay Reference Hay, Binder, Rhodes and Rockman2009, 59; see also Steinmo Reference Steinmo, Della Porta and Keating2008, 123–126; Schmidt Reference Schmidt2010: 2). This broader construct suggests that the way that institutions constrain, shape and enable policy is by providing a “logic of appropriateness” which guides behaviour and means that “the important institutions (rules) are social norms that govern everyday life and social interaction” (Steinmo Reference Steinmo, Della Porta and Keating2008, 126).
The latter views remain contested as overbroad, with Streeck and Thelen cautioning against a reliance on a definition of institutions as “too broad to be meaningful”, and rejecting “shared cognitive templates” as the basis for a definition of institutions, because they elide the gap between institutions and behaviour, thereby rendering invisible “conflict over competing interpretations that could be explored as a source of change” (Streeck and Thelen Reference Streeck, Thelen, Streeck and Thelen2005, 11l; see also Powell and DiMaggio Reference Powell and Di Maggio1991, 15; Lowndes and Roberts Reference Lowndes and Roberts2013, 32).
These two conceptualisations of norms acting in institution-like ways are very different. For the purposes of the argument here, I will differentiate between the narrower definition of norms acting in institution-like ways, through formal and informal rules and sanctions, and the broader definition of norms acting in institution-like ways as cultural and cognitive frames. I will show how these two types of institutionalisation of the same norm operate differently, and in fact can operate in contradiction with one another, such that the broad level masks policy change at the narrow level.
Research design
This is a qualitative study of policy change in three countries: the US, the UK and Australia. These countries provide a good basis for comparison, since they are all broadly similar liberal democratic states, in that they uphold the rule of law and in them free speech is a culturally and legally accepted norm that has been instantiated both in broad, cultural frames and in narrow, formal rules. At the broad level, all three countries claim to adhere to a socially and culturally framed principle of freedom of speech, as evidenced both by public discourse and by the importance these countries place on freedom of speech in political culture (Barendt Reference Barendt2007, 39–55, 71–73; Saunders Reference Saunders2017, 7–10, 13–15). In all three countries, government leaders routinely emphasise the importance of free speech. For example, US President George W. Bush declared, “America will always stand firm for … free speech” (Reference Bush2002a), while Attorney General Ashcroft declared “free speech and open inquiry … [to be] the bedrock upon which freedom stands” (Ashcroft Reference Ashcroft2002), and that his nation was committed to “supporting freedom of speech” (Ashcroft Reference Ashcroft2003a). In the UK, in 2005 Home Secretary Charles Clarke remarked that, “the United Kingdom and the United States are societies which value and build free speech and freedom of expression” (Reference Clarke2005) and Prime Minister Gordon Brown described “rights for the public expression of dissent” as an “enduring ideal” (Brown Reference Brown2007). Australian Prime Minister John Howard expressed the view that, “countries like Australia … value freedom” (Reference Howard2002a), and that Australian attitudes and values “clearly include free and open expression” (Reference Howard2002b; for further examples see Howard cited in Hudson Reference Hudson2002; DPMC 2008, 310, 322–324).
I focus in this article on tracing how policy changes occurred in relation to freedom of speech in the US, the UK and Australia in the decade after the 2001 terrorist attacks, and in the context of national security debates. This is because significant policy change occurred during this time resets the previously accepted boundaries of freedom of speech in relation to speech considered harmful to national security. This investigation of policy change triangulates data from three sources.
First, the study identified changes to the parameters of free speech by collating all new or amended national security legislation that posited speech as harmful to national security, and restricted it, in the period 2001–2011. Some of the most significant changes to the accepted parameters of free speech occurred in this area, in which speech-based activities were restricted by criminalising speech that was considered in some way to “assist” terrorism. This included (as will be shown below) the criminalisation of some speech that was far removed from creating a tangible terrorist risk.
Second, the study involved analysis of national-security-related speechesFootnote 1 given by key policy makers: Presidents, Prime Ministers,Footnote 2 Attorneys General and Home SecretariesFootnote 3 in the period from September 2001 to September 2011. These speeches were subjected to qualitative discourse analysis,Footnote 4 in which the researcher interprets and categorises language use to determine the communicative strategies used by authoritative speakers to legitimise their policy proposals and delegitimise opposition to them (van Dijk Reference van Dijk and van Dijk1997, 2, 10–11; Chilton and Schäffner Reference Chilton, Schäffner and van Dijk2011, 311–312). The analysis sought to describe the characteristics of the communications (Breuning Reference Breuning, Ishiyama and Breuning2011, 491–492) by drawing out common themes used by the speakers. The purpose of this discourse analysis is to trace the justifications posited by key policy makers for statutory changes that impacted on freedom of speech.
Third, documents from other institutions and organisations engaged in debates about national-security-related policy change that impacted on freedom of speech were analysed. This includes formal institutional actors such as courts in which statutory provisions have been challenged in judicial proceedings, the UK Joint Committee on Human Rights’ (JCHR’s) reports on national security-related legislation and its impact on human rights from 2001 to 2011 and government responses to those reports, relevant reports by the Australian Parliamentary Joint Committee on Intelligence and Security and the Senate Legal and Constitutional Affairs Committee, and reports by civil society organisations such as Article 19, Human Rights Watch and Amnesty International that intervened in the public debate around policy change. Reports from these sources were analysed to trace what policy changes were initially proposed, the grounds on which opposition was expressed to those changes, the extent of the resulting policy change compared with the original proposals and the ways in which the policy change as proposed and enacted impacted on freedom of speech.
Empirical setting
In the US, free speech is regarded as a more important political value compared with other countries (e.g. Cohn Reference Cohn2012). In the UK a historical common law protection for freedom of speech provides evidence of a historical, cultural commitment to free speech (Rosenfeld Reference Rosenfeld, Herz and Molnar2012, 262), as does the country’s participation in the European Convention on Human Rights, and the affirmation of the importance of free speech by, among other institutions, the parliamentary JCHR (e.g. JCHR 2012, 61). Australia inherited its common law, and therefore a historical, cultural recognition of free speech, from the UK and survey results show ongoing public support for freedom of speech (Gelber Reference Gelber2011, 25).
Although each country adheres to the same norm, there is considerable difference between them in the instantiation of the norm at the level of narrow, formal rules. The US contains the strongest legal protection of freedom of speech globally in the form of the First Amendment, which prohibits Congress from making laws abridging the freedom of speech. Its exceptional (Schauer Reference Schauer and Ignatieff2005) provision protects freedom of speech in public discourse (Post Reference Post2011; Weinstein Reference Weinstein2011) from government regulation to a very high degree,Footnote 5 with a few narrowly delineated exceptions including true threats, fighting words,Footnote 6 and some libel and obscenity (Barendt Reference Barendt2007, 48). The UK’s Human Rights Act 1998 gives effect to Article 10 of the European Convention on Human Rights, meaning that freedom of speech has been explicitly protected in statutory law in that country since the Act came into force in 2000. Although this was a change from the previous common-law basis on which freedom of speech had been protected, the scale of policy change is regarded as insignificant, indeed not much more than “cosmetic” (Barendt Reference Barendt2007, 39). Australia has augmented its common law protection for freedom of speech since the development in 1992Footnote 7 by the High Court of Australia of a constitutionally implied freedom of political communication, derived from the implications of the constitutionally prescribed system of representative and responsible government (Gelber Reference Gelber2016, 23–25). This doctrine, however, has not operated strongly to protect freedom of speech because it still permits the valid restriction of political speech, as long as the purpose of the law that does so, and the means adopted to do so, are legitimate, do not “adversely impinge upon the functioning of the system of representative government” and are “reasonably appropriate and adapted to advance” the law’s purpose, in terms of its suitability, its necessity and the adequacy of its balance.Footnote 8
Therefore, in all three countries freedom of speech is routinely defended as a core component of the extant political order, as a broad, culturally framed norm. It is a norm by which governments agree they should abide. It appears to impose expectations of appropriate behaviour on actors, as evidenced by policymakers asserting its importance. Additionally, all three countries have different narrow, formal rulesFootnote 9 around free speech and set out enforceable laws in relation to which there are legally delimited circumstances under which the norm can be restricted, along with sanctions for noncompliance. This suggests that free speech in these countries survives threshold definitional tests of being an institution, in both the broad and narrow senses outlined above. However, it does not yet show how free speech acts in institution-like ways, or what this reveals about how norms-based institutions can constrain, shape and enable policy change.
Policy change
The international context for policy change after September 2001 was important in driving policy change. There was significant pressure on governments from the international community to act decisively to counter terrorism, including the passage by the United Nations Security Council of Resolution 1373 on 28 September 2001, which called on member states to take action to counter the risks of terrorism. Further United Nations resolutions, treaties and a counterterrorism committee were devoted for this purpose.Footnote 10 This gave an imperative and justification to policy change that was leveraged by policymakers, who dedicated significant new financial and human resources to achieving this change (Roach Reference Roach2011, 1–5).
Importantly, the discourse of key policymakers in the three countries supported this urgent and pressing need for policy reform in ways that were strongly consistent. In all three countries, political leaders’ speeches demonstrated very similar justifications for significant policy change, which resulted in restricting freedom of speech in ways that had not previously been considered acceptable. The examples given below are representative of their discourse. These justifications included three key elements.
The first element was the positing of the events of 9/11 as a “war” that was qualitatively different from any situation that had preceded it. For example, US President George W. Bush and Attorney General Ashcroft both regularly described the situation as “a new war” (e.g. Ashcroft Reference Ashcroft2001a; Bush Reference Bush2001a), in which “the nature of the war is different” (e.g. Bush Reference Bush2002b). The UK Prime Minister Blair described it as a “new type of war” (Blair Reference Blair2002a), and the Home Secretary as a “war against humanity” (Blunkett Reference Blunkett2004). The Australian Prime Minister described it as “a new kind of war” (Howard Reference Howard2001a) and the Attorney General as “not a traditional war” (Williams Reference Williams2002a).
Second, the new threat was posited as requiring a different kind of policy response, including the development of qualitatively different policies from those that had previously existed. The US Attorney General stated, “we need a new way of doing business” (Ashcroft Reference Ashcroft2001b), and the US President that the situation “require[s] a new way of thinking” (Bush Reference Bush2001b; Reference Bush2005); it is time to “think and act differently” (Bush Reference Bush2002c). Bolstered by “new authority” (Bush Reference Bush2001c), the US President and Attorney General called for “unprecedented” (Bush Reference Bush2003a) policies that would “push to the limits of the law” and stimulate debate over “whether the [legal] lines themselves should be redrawn” (Mukasey Reference Mukasey2008). The UK Prime Minister famously declared, “the rules of the game are changing” (Blair Reference Blair2005a), and that, “we can’t tackle terrorists by the rules of the game we have now” (Blair Reference Blair2005b). But well before this he had already suggested that this was a “new situation … without parallel” (Blair Reference Blair2001a), one that meant “we need to rethink dramatically the scale and nature” of remedial action (Blair Reference Blair2001b). He asserted that, “new threats need new measures” (Blair Reference Blair2002a) and that “we do need to contemplate things that maybe a few decades ago we wouldn’t have” (Blair Reference Blair2003). Indeed, the new situation meant “our whole thinking should change” (Blair Reference Blair2004). Similarly, the Australian Prime Minister and Attorney General regarded what was needed under these circumstances as “vastly different” (Williams Reference Williams2002b; Howard Reference Howard2002c). The Prime Minister and Attorney General argued governments needed to be able to do “things we ordinarily would not have done” (Howard Reference Howard2001b), even where those things “need to go further than the current law” (Howard Reference Howard2005) and were “extraordinary” (Williams Reference Williams2002c).
Third, and in a vital component of the overall messaging, key policymakers posited the need to ensure safety and security first, in order that liberties could be enjoyed later. US President George W. Bush declared, “my most important job is the security and safety of the American people” (Bush Reference Bush2002d), and US Attorney General Gonzales that, “without security, there can be no real freedom” (Gonzales Reference Gonzales2005). The UK Prime Minister Gordon Brown affirmed that, “our primary duty – abiding obligation – is safety” (Brown Reference Brown2008), and Home Secretary David Blunkett that, “individual freedom rests on internal and international order and stability” (Blunkett Reference Blunkett2003). The Australian Attorney-General opined that, “our most important duty … [is] protecting our country and our people” (Ruddock Reference Ruddock2004), and Prime Minister Howard that, “human security is the foundation of what it means to be free” (Howard Reference Howard2004).
There was, therefore, a commonality of purpose as expressed in speeches by these key policymakers, one that demonstrated strong consistency between the three countries. The common purpose expressed was the need to recognise the novelty of the situation, a need and intention to reset, and push the boundaries of, existing policy to deal with the new threat and the need in so doing to ensure security as the primary duty of government. The identification of this commonality of purpose is neither conspiratorial nor coincidental – these governments cooperate closely on the development of counterterrorism measures and routinely share counterterrorism intelligence (Roach Reference Roach2011: 18; Walker Reference Walker2011). Having established the commonalities of purpose and intentions among key policymakers, I move now to consider specific policy changes that occurred in each country.
United States
In the US, freedom of speech is protected with a very high degree and much speech, even speech that is considered harmful, is not regulable. Since the 1960s the jurisprudence of the Supreme Court has strongly enforced a presumption against the validity of content-based, and viewpoint-based, laws that restrict freedom of speech, subjecting them to strict scrutiny to assess their validity.Footnote 11 This has rendered speech restrictions invalid, including a city ordinance prohibiting the dissemination of material promoting racial or religious hatred,Footnote 12 and an ordinance prohibiting placing objects that arouse anger or alarm on the basis of race, religion or gender.Footnote 13 Although in sedition cases up to the 1950s the Supreme Court had accepted a legislature’s assessment, “that speech of a certain character is per se dangerous”,Footnote 14 by the 1960s this was no longer the case (Barendt Reference Barendt2007, 158, 165) and legislatures were not permitted to designate speech as inherently dangerous based on its content or viewpoint.
After the 2001 terrorist attacks the US government moved swiftly to legislate to try to prevent further attacks, with the Attorney General asking staff in the Department of Justice to draft “an aggressive set of legislative changes” (Brill Reference Brill2003, 52) on the same day the attacks occurred. Although the Attorney General came to express the view that “extolling” amounted to support for terrorism (Ashcroft Reference Ashcroft2003b), it was clear that the extant constraints of the First Amendment prevented the introduction of explicitly speech-limiting provisions, such as extolling, glorification or encouragement of terrorism (Roach Reference Roach2011, 227).
Other policy options were pursued that were both possible within, and rewrote, the extant constraints. The primary tool was federal criminal law prohibiting providing “material support” to a designated foreign terrorist organisation, which was amended in s805 of the PATRIOT Act in 2001. The change expanded the definition of “material support” from what one might normally consider to constitute material support such as finances, lodging, training, documentation, equipment, facilities, weapons, personnel and physical assets to include “expert advice and assistance”.Footnote 15 Following public criticism of, and legal challenges to, the lack of clarity in the meaning of “expert advice and assistance” (De Rosa Reference De Rosa, Baker and Kavanagh2005, 143; Lombardo et al. Reference Lombardo, Buwalda and Lyman2006, 239), Congress amended the provision in 2004 to define it as “advice or assistance derived from scientific, technical, or other specialized knowledge”,Footnote 16 and to require that a person giving advice know that the organisation to which they are giving advice is a designated foreign terrorist organisation.Footnote 17
A legal challenge was launched to the speech-based components of the material support provision by the Humanitarian Law Project (HLP), a nongovernment organisation that had worked with the Kurdistan Workers Party and Tamil organisations to educate them in how to pursue their ends using peaceful and legal means. The HLP questioned whether their work could be caught up by the material support provision, even though their advice was designed to counter the risk of terrorism and violence. This case can be seen as an attempt to reinforce the preexisting, narrow institutionalisation of free speech through a judicial challenge to the new policy.
Before the courts, the government argued that the speech engaged in by the HLP should be considered to constitute material support. It argued that, like all other forms of support to terrorist organisations, expert advice and assistance should be considered fungible, by which it meant providing such advice might free up resources elsewhere in the organisation that could be used to pursue terrorism. In 2010 the US Supreme Court ruled in the case.Footnote 18 Unexpectedly in the eyes of many, it upheld the government’s argument and found that the material support provision facilitated the “compelling government interest” of combatting terrorism and employed a balancing methodology. The result was that the Supreme Court did not provide First Amendment protection to purely speech-based advice, even where that advice was designed to dissuade groups from engaging in terrorism (Cole Reference Cole2003a, 61; Reference Cole2003b, 10–11; 2005, 145–146). This altered preexisting understandings of the contours of First Amendment jurisprudence in relation to speech alleged to harm national security.
United Kingdom
In the UK freedom of speech is also strongly protected, but the UK has enacted legally permissible restrictions on the free speech right, which are far wider than those permitted in the US (Bleich Reference Bleich2011, 5–7). They are not unlimited; however, both before and after the implementation of the Human Rights Act 1998, courts in the UK have protected freedom of speech in the context of libel, censorship, public order offences and press freedom (Barendt Reference Barendt2007, 41, 46–47).
The constraints of the Human Rights Act were visible in the options chosen by policy agents in successive cycles of policymaking. Since explicitly speech-limiting provisions were allowed, they were pursued as part of the prevention-of-terrorism agenda. In fact in 2000,Footnote 19 prior to the 2001 terrorist attacks, the government had made it possible to proscribe organisations that promoted or encouraged terrorism, with membership of such organisations also a criminal offence (Walker Reference Walker2011, 357–358), an option that was not possible in the US due to the constraints of the First Amendment. At the time this was considered to be a comprehensive updating to the UK’s counterterrorism laws (Walker Reference Walker2006, 1138). However in 2001,Footnote 20 and again in 2006Footnote 21 following the London terrorist attacks, the government enacted new counterterrorism laws, including in the latter iteration by expanding the proscription provision to define an organisation as promoting or encouraging terrorism where it engaged in the “glorification” of terrorism, with glorification defined as including “any form of praise or celebration”. At the same time in 2006, the encouragement of terrorism was made a discrete criminal offence.
The 2006 legislative changes eventuated after a draft bill was released in 2005, which was heavily criticised, including in the parliamentary JCHR and in public debate. Earlier drafts of the provisions were more far-reaching in their impact on freedom of speech than the finally enacted version, as they sought to create discrete offences of encouragement of terrorism, and of glorification of terrorism (Amnesty International 2005; Article 19 2005, 2; Human Rights Watch 2005;Metcalfe Reference Metcalfe, Reed and Dumper2012, 153). The originally drafted provisions did not require intent that a terrorist act occur as a result of the encouragement, were vague in the definition of glorification and were argued to be overbroad in the types of speech they could capture (JCHR 2006, 16–26). This indicates a desire on the part of policymakers to reach further than extant constraints of the narrow institutionalisation of free speech, and other institutions, permitted. These debates led to changes being made to the proposed legislation before it was enacted in order to better protect freedom of speech, but it remains the case that new provisions introduced significant new constraints on speech-based activities in the context of national security.
Australia
In Australia the protection of freedom of speech is the weakest of all three countries, and a range of speech restrictions that would be impermissible in the US has survived constitutional scrutiny (Gelber Reference Gelber2016, 23–25), including the exclusion of protestors from a duck-shooting area,Footnote 22 a prohibition on the use of threatening, abusive or insulting words in public,Footnote 23 and requiring a person to obtain a permit to preach or distribute material in a pedestrian mall.Footnote 24 Given the weak protections afforded freedom of speech in the legal and regulatory framework, the policy options pursued were the most far-reaching of all three countries. In this context, it is highly interesting to note that other institutional factors, including parliamentary opposition, and public debate played a weak role and had relatively little traction in resisting radical policy options when they were introduced.
Australia introduced laws criminally proscribing organisations that advocate terrorism,Footnote 25 with advocacy being broadly defined as including “directly or indirectly” counselling, urging or instructing, or directly praising where that substantiallyFootnote 26 risks “leading a person (regardless of his or her age or any mental impairment that the person might suffer) to engage in a terrorist act”. In addition to the breadth of this phrasing, there is no requirement that a person intends that terrorism occurs as a result of their advocacy, and members of organisations in which one person makes a comment able to be captured by the legislation could become criminally liable (SLCAC 2005, 128–129). Although there was considerable community opposition to, and an inquiry by the Senate Legal and Constitutional Affairs committee into, the draft laws (Chong et. al. Reference Chong, Emerton, Kadous, Petitt, Sempill, Sentas, Stratton and Tham2005; SLCAC 2005), no amendments were made between the proposal and the enactment of the laws. In 2014, the government also introduced a new, discrete offence of advocating terrorism (Brandis Reference Brandis2014).
Additionally, the Australian government adopted a policy of (re)introducing “sedition” provisions, trying to turn them to the task of protecting the community against speech that might result in terrorism by criminalising urging violence between groups defined by race, nationality, religion or political opinion. The same parliamentary committee report noted “overwhelming opposition” from the community to these laws and strong public criticism (SLCAC 2005, 1, 6, 86–89; Marr Reference Marr2007, 66; Gelber Reference Gelber2009, 271–278). But again, the laws were enacted with little change. They were amended in 2010Footnote 27 inter alia to rename them “urging violence” offences and to expand their remit both to groups and to members of groups. The reach of these laws is very wide, and the terms in them vague and broad. Therefore in Australia, the country with the weakest narrow institutionalisation of free speech, constraints on policy change were minimal and far-reaching restrictions that would likely not have been possible in the other two countries were implemented.
Constraint, shaping and enabling
Across the three countries, the processes outlined here evince constraints on, and the shaping and enabling of, policy change that derived from the instantiation of free speech in political culture, legal rules and judicial interpretations. They also evince the successful rewriting of the parameters of free speech through the introduction of significant new criminal provisions limiting speech that was perceived to be a risk to national security, even where the provisions were overbroad and overreached that mandate. In the US a new speech-limiting provision was designed that would survive the extant First Amendment strictures on freedom of speech, and that restricts speech far beyond that which supports terrorism. The extant framework constrained the options available for policy change, in so far as it prevented the introduction of an explicit offence of encouraging, promoting or extolling terrorism. New policy was both constrained and enabled by other institutions; it was challenged by citizens in the courts, with initially inconsistent results in lower court judgements.Footnote 28 Eventually, the Supreme Court confirmed that the new speech limitation was valid.
In the UK debate on exactly where it was appropriate to draw the line on freedom of speech and limiting harmful speech on national security grounds also featured. Policy proposals tried to push the limits of freedom of speech further than would eventually be achieved, but they were constrained in doing so by other elements of the institutional context. The 2005 attempt to redraw the line on permissible versus regulable speech was partially pushed back by human rights advocates, the JCHR and in parliamentary debate. In the end, however, policy change still resulted in significant new, criminal speech-limiting provisions that redrew the boundaries of acceptable limits on freedom of speech. In Australia a weak, narrow institutionalisation of freedom of speech played a role in policy change, as did other aspects of the institutional context in which that change occurred. These included the fact that parliamentary debates and extraparliamentary opposition had very little traction in achieving pushback against quite radical proposals, which meant significant policy change was able to be achieved that reset the accepted parameters of freedom of speech in capacious ways. This story of policy change shows how the proposed changes to policy, and opposition to and criticism of these changes, both reflected the institutionalisation of the free speech norm in each country and demonstrated strong interactions with other institutions.
Discussion
Did free speech act in institution-like ways in this account of policy change, and if so what does this reveal about how norms act in institution-like ways? Earlier, I suggested that a norm can act in institution-like ways if and when that norm is instantiated in formal and informal ways, is relatively resilient to changes desired by individual, idiosyncratic preferences and is instituted in such a way that there are sanctions for deviation from the norm. These criteria appear to have been fulfilled. The extant parameters of free speech played a material role in constraining, shaping and enabling the options available for policy change. They also played a material role in framing and shaping the kinds of pushback that took place against the policy change, and arguably also the success or otherwise of that pushback.
As demonstrated, in each country the broad institutionalisation of free speech meant rhetorical commitments to free speech were maintained. At the same time the narrow institutionalisation of free speech materially affected the options for policy change, even in a context where policymakers had shown a determination to reset policy boundaries in response to what they perceived as a new threat. Policy change narrowed the formal rules and institution of free speech considerably, by expanding the criminal prohibition of speech perceived to be harmful to security. This is the case even though some of the policies were written so broadly as to capture speech that presented only a tangential, if any, risk of terrorism.
Having answered the question of whether free speech can be conceptualised as acting in institution-like ways in the affirmative, the related question that arises is what does conceptualising free speech as an institution illuminate about processes of policy change? There are three observations to be made. First, free speech itself – the norm, the broad institution and the narrow institution – has been changed in processes of interaction with other institutions, including legislatures, courts, parliamentary committees and political discourse. This complicates how we understand norms acting in institution-like ways and shows they can simultaneously act as vectors and as objects of change. The story of policy change outlined here is both one of institutional constraints, shaping and enabling, and one of changes resulting to the norm and therefore also to the institution of free speech in this context.
Although constraints were evident, policy change resulted in the successful reconstruction of the parameters of free speech in the context of national security. The policy change took on a pace, direction and continuity that reached well beyond the terrorist events that temporally marked the onset of these changes.Footnote 29 This change occurred in an international context that posited the changes as necessary and appropriate, and domestic contexts in which policymakers consistently posited the change as essential, necessary, important and appropriate. This context meant policymakers attempted to undertake far-reaching reform, even in the context of the institutionalisation of free speech that suggested in all three countries (most strongly in the US, to some degree in the UK and weakly in Australia) that their reforms might not be acceptable. Policymakers interacted with the “institutional and wider structural contexts” (Bell Reference Bell2011, 884) to achieve significant policy change. The story here was not, either in a temporal or a causal sense, limited to responding to the events of 11 September 2001; it encompasses long-term shifts in the institutionalisation of free speech in the context of national security. This is commensurate with New Institutionalist literature that shows that, and how, situated agents respond to resources, challenges and opportunities in the institutional contexts in which they find themselves (Bell Reference Bell2011, 884; Bell and Feng Reference Bell and Feng2014).
The second observation is the disjunct, indeed contradiction, between the broad and narrow institutionalisations of the same norm. At the broad level, free speech acted as a cultural norm, a cognitive frame that guided the behaviour of those involved in the policy debates. This is visible in policymakers’ statements in defence of the value of free speech, in the members of parliamentary committees who sought to reinforce a stronger protection of free speech against attempts to restrict it, and in civil society actors’ interventions emphasising the need to protect free speech against encroachment. However, at the narrow level significant change occurred that undermined the protection of free speech. In this context, the adherence to freedom of speech at the broad level by policymakers acted as a rhetorical framing device that enabled them to claim a commitment to free speech: the broad level of a cultural script provided a logic of appropriateness that was powerful enough that governments felt compelled to claim a commitment to it. But those governments simultaneously undertook policy change that significantly altered the narrow, formal institutional rules in each country, to restrict freedom of speech more than had been the case before the policy changes took place. This means that the broad cultural script acted rhetorically to frame policymakers’ capacity to undertake change, and at the very same time it masked underlying policy change that undermined, and even contradicted, the norm. This means the broad institutionalisation of a norm can obfuscate, rather than illuminate, policy change at the narrow, formal level; it can mask detailed policy change which contravenes the norm it instantiates, by allowing policymakers to claim cultural adherence to the norm while simultaneously violating it.
This expands our understanding of how norms-based institutions operate, by illuminating differences between the effects of the broad and narrow institutionalisation of a norm. This conclusion differs from Thelen’s view that the concept of cultural scripts “obscures political struggles among competing scripts” (Thelen Reference Thelen1999, 384), and Streeck and Thelen’s caution (Reference Streeck, Thelen, Streeck and Thelen2005, 11) that defining institutions in a way that is too broad to be meaningful can elide the gap between institutions and behaviour. Examining the two levels of institutions has revealed that the broad institution was relied upon in a way that masked contradictory institutional change at the narrow level. This suggests that norms can act institutionally at the broad level differently from, and even in contradiction to, their operation at the narrow level in accounting for policy change. They can serve as a mask for changes to formal rules that undermine the norm. At the broad level policymakers can abide by the logics of appropriateness to which a norm gives rise and fulfil the conditions for the norm to be acting in institution-like ways, while at the same time undertaking policy change that may be undermining that very same logic of appropriateness.
Finally, this analysis produces an observation, which is related to the existence of a wide variety in the formal rules that institutionalise freedom of speech. This is the case among these otherwise similar polities, and even more so when free speech is viewed globally. The analysis here suggests that one can understand the difference between the institutionalisation of the norm of freedom of speech at the broad level and at the narrow level, and more easily accept that the latter can – and indeed routinely does – differ from the former in terms of the ways in which the protection of freedom of speech is instantiated, because the institutionalisation of a norm will differ from polity to polity. Relatedly, one can accept that freedom of speech at the broad level tells us very little about the rules by which the protection or otherwise of freedom of speech is achieved on the ground. This helps to understand how countries with different approaches can all claim to be defending the same norm, even though the narrow institutionalisation of that norm differs widely. We no longer need to debate endlessly whether or not freedom of speech is protected to the “right” degree in a particular country, based on philosophical, normative understandings of where the boundaries should lie. Instead, freedom of speech as a set of practices with rules and sanctions for noncompliance can be seen as resulting from differentiated historical pathways, which combine to constrain, shape and enable some (but not all) potential options for policy change. Complexity and variation in speech regulation can be understood as products of the institutions at play, and the to-be-expected variability in the institutionalisation of a norm.
Conclusion
In this article I have examined whether the norm of free speech, a norm that is highly valued in the three countries under study, has acted in institution-like ways, understood as constraining, shaping and enabling policy change. The first original contribution of this article, therefore, is to the literature on norms as institutions because it is the first time free speech has been considered in this light.
In conducting the analysis, I differentiated between a narrow definition of when norms can act in institution-like ways, through formal and informal rules and sanctions, and a broader definition of norms acting as cultural and cognitive frames. I examined policy change between 2001 and 2011 in the USA, the UK and Australia in the context of national security laws and their impact on freedom of speech. An important element of this policy change was the justifying discourse of key political leaders, who posited the 2001 terrorist attacks as ushering in a qualitatively different kind of war, which required a new policy response that reached further than had previously been permitted, and was necessary to secure public safety.
The policy change achieved was considerable. In the USA, in spite of the constraints of the First Amendment, new provisions were introduced that criminally prohibit purely speech-based advice, even when that advice is designed to dissuade people from engaging in terrorism. In the UK, new criminal provisions prohibit the encouragement of terrorism based on a capacious conception of encouragement. In Australia, far reaching limitations on speech have been introduced, even where they have little or no connection to substantive terrorist threats. In all three cases, the instantiation of free speech in political culture, legal rules and judicial interpretations constrained the options available for policy change. Nevertheless, significant restrictions were introduced, which occurred in the context of each government maintaining ongoing rhetorical commitment to the norm they were reshaping.
The analysis has shown that, and how, free speech acted in institution-like ways, including showing that norms-as-institutions can operate both as vectors and as objects of policy change. It revealed that there are differences, contradictions even, in how institutionalised norms operate to constrain, shape and enable policy change at a broad level and at a narrow level. This implies that, when analysing norms acting in institution-like ways, greater attention should be paid to these differences and what they can reveal about policy change.
Finally, the analysis has shown that differences in the instantiation of a norm at the broad and narrow levels are to be expected given the different political contexts in which the norm is mediated. Indeed, the institutionalisation of a norm helps to explain variability in application of the same norm across different cases, variability that may otherwise be difficult to comprehend. The logical questions that arise from this study are whether these insights apply to the institutionalisation of all norms, or perhaps only to certain categories of norms or perhaps only under certain conditions of policy change. Since there is no room to explore these questions here, they are topics for future study.
Author ORCIDs
Katharine Gelber, 0000-0002-0239-3552
Acknowledgements
This study was funded by the Australian Research Council (FT110100114) and granted Ethics approval by the University of Queensland (2011001333). The author wishes to thank Stephen Bell, Alastair Stark, Andrew Phillips, Chris Reus-Smit, Nicole George, Alexander Brown and Susan Brison for helpful discussions on the ideas in this article, and Ryan Smith and Danielle Le Poidevin for research assistance in the collection of data. The author wishes to thank the journal’s referees for very helpful feedback.
Data Availability Statement
This study does not employ statistical methods and no replication materials are available.
