Silencing and freedom of speech in UK higher education

Freedom of speech in universities is currently an issue of widespread concern and debate. Recent empirical findings in the UK shed some light on whether speech is unduly restricted in the university, but it suffers from two limitations. First, the results appear contradictory. Some studies show that the issue of free speech is overblown by media reportage, whilst others track serious concerns about free speech arising from certain university policies. Second, the findings exclude important issues concerning restrictions to speech on campus that fall outside the traditional debate around violations to free speech rights. This is particularly the case when certain voices are excluded from important policy conversations, and in issues around diversifying the curriculum. This article overcomes these two limitations by developing a novel conceptual framework within which to situate current debates concerning speech-related matters in universities. It does so by developing a taxonomy around the concept of ‘silencing’. It then considers the current issue of speech matters in universities within this framework to determine whether, and to what extent, speech is indeed unduly restricted, and where this is a concern for free speech violations, and where it falls outside this issue.


Freedom of speech in UK universities
Freedom of speech is currently a highly controversial issue in public and political debate. It is especially prominent within universities given divisive 'safe spaces' and 'no-platforming' activities, and apparent limitations to academic freedom (Lackey, 2018). Amidst the widespread media attention on, and public outcry concerning, freedom of speech in the last few years, several reports and empirical studies have been produced that seek to identify whether freedom of speech is unduly constrained in universities in the UK, and if it is, what the extent of restrictions to freedom of speech has been.
The outcomes of these studies pose several problems, most prominently that their findings appear to offer conflicting outcomes. For instance, in a survey conducted in 2019 by YouGov for UK-based religion and society thinktank Theos, it was reported that 52% of the 2,041 respondents said that they believe freedom of speech to be 'under threat' in universities (Perfect et al., 2019a, p. 10). However, Theos's subsequent report on freedom of speech insists that such widespread perceptions are unfounded.
They claim that 'the widespread public concerns identified in the Theos 2019 polling -that freedom of speech is under threat in universities (52% agree) . . . are exaggerated or overblown' (Perfect et al., 2019a, p. 143). This report takes its lead from the UK government's Joint Committee on Human Rights (JCHR) in 2018, which maintained that perceptions about a crisis of free speech is due to misleading media reportage. The JCHR argued that 'press accounts of widespread suppression of free speech are clearly out of kilter with reality . . . the narrative that "censorious students" have created a "free speech crisis" in universities has been exaggerated'. On the basis of their review, the JCHR concluded that they 'did not find the wholesale censorship of debate in universities which media coverage has suggested'. 1 These conclusions from Theos and the JCHR, however, stand in tension with other surveys and interviews amongst students. For instance, in 2016, the Higher Education Policy Institute-a thinktank dedicated to research in UK higher educationconducted a survey of 1,006 full-time undergraduates at UK universities. They found that only 41% of students reported feeling 'completely free' to express their opinions and political views openly and without any restriction (Hillman, 2016, p. 7). Moreover, in a recent study of the experiences of Muslim students in UK universities, over half of the 253 interviewees and focus group participants believed that state securitisation measures discriminated against Muslims because of their religion (Scott-Baumann & Perfect, 2020). In particular, the government's 'Prevent Duty' policy has caused Muslim students to feel like objects of suspicion, and to 'self-censor' their speech. 2 So, on the one hand, some research outcomes downplay the issue of freedom of speech in higher education, whilst others maintain that the issue is serious and pervasive. One aim of this article is to resolve this apparent tension. I argue that the fact that these studies appear to offer conflicting outcomes is in part due to their coarsegrained approach to the issue. The topic of freedom of speech covers a range of distinct concerns that are sometimes placed together. Legal prohibitions on the public use of hate speech, campaigns against giving platforms to controversial views, and students feeling prejudiced against for sharing their beliefs and opinions can all fall within the general category of restrictions to speech, but are entirely distinct issues. Moreover, some matters relevant to this debate fall outside the notion of freedom of speech, but are nonetheless relevant to current matters of speech on campus. This is the case when, for instance, a non-inclusive syllabus on certain topics alienates students from the discussion, leading to disengagement and self-quieting. Though this might be relevant to the current concerns around speech in higher education, it needn't be construed as either a violation or justified restriction of free speech rights.
This article aims to facilitate better understanding of the issues concerning speech in universities. In order to do so, it will provide a conceptual framework within which to identify and evaluate the issues highlighted by some recent survey and interview data. Rather than working strictly within the language of freedom speech, this article develops a framework in terms of the more nuanced concept of silencing. Recent research on silencing has generated a vibrant literature at the intersections of philosophy of language and social and political philosophy, and these ideas provide a fresh basis upon which to analyse the issues at hand.
Fitting the current discussion on free speech in universities within a silencing framework is useful in two respects. First, since there are numerous varieties of silencing, we can distinguish between different issues around free speech and map these onto a taxonomy that is fine-grained. This prevents overlap of discussion and makes visible the separateness of the issues. Second, silencing shows where there are problems that go beyond the current focus on free speech. We can show that whilst all instances of restrictions to freedom of speech involve silencing, some instances of silencing do not involve restrictions to freedom of speech, which are nevertheless relevant to current issues concerning speech in higher education.
The benefit of analysing empirical work concerning freedom of speech within a silencing framework is bilateral. For empirical research, we can provide a way of distinguishing the ways in which silencing occurs and does not occur, and offer a means of determining the extent to which such silencing is morally problematic. For philosophical theory, results from surveys and interviews can be used to demonstrate live issues of silencing, and to help expand upon the overall taxonomy of its different varieties.
Where this article addresses the issue of free speech rights, it will mainly discuss examples concerning students on campus, drawing from surveys and interviews conducted with students. There will be very limited discussion of the connected issue of academic freedom. The relation between the two, which I explain in detail, is that academic freedom is a restricted form of freedom of speech, applying to 'a limited class of communicative activities, coming from a limited class of speakers' (Simpson, 2020, p. 291). Whilst freedom of speech covers all speakers within a designated legal jurisdiction, academic freedom concerns the issue of free speech for activities undertaken by professional academics in a university context, especially the content of research and teaching. Nevertheless, the silencing framework I develop can be just as readily applied to academic freedom as to free speech amongst students.
The article proceeds as follows. First, I give an account of freedom of speech and outline its legal and policy context within UK higher education. Then, I offer a taxonomy of different kinds of silencing, which is elaborated upon by drawing on examples from recent work in sociology. Following that, the extent of the issue concerning silencing, and its bearing on freedom of speech, will be explored. It is argued that some prominent kinds of silencing have indeed been overinflated by media reports, but that other forms of silencing are becoming more widespread, and are advanced in part due to recent university and governmental policies.
The legal and policy context in the UK One way to conceive of freedom in general is in terms of Isaiah Berlin's conception of negative liberty: I am normally said to be free to the degree to which no man or body of men interferes with my activity. Political liberty in this sense is simply the area within which a man can act unobstructed by others. (Berlin, 2002(Berlin, [1969 For Berlin, freedom is restricted in a political sense-in the sense involving civil rights-when people or institutions obstruct our actions. We are free, then, when we Silencing and freedom of speech in UK higher education 3 are unobstructed. Freedom of speech, then, is the lack of any hindrance to speaking on the part of other people, or other institutions, including any body of law in accordance to which someone may be punished for speaking. The notion of free speech as a civil right, with attendant implications for law and policy, has both a scope and a guarantee. The scope of free speech concerns whether there are any laws under which there is threat of state-sanctioned punishment for saying a particular thing. For instance, in a particular state, if there are no laws under which people can be punished for any of their speech, other than when they discriminate against others, then the scope of the freedom those people have extends to saying anything, except discriminating against others. The guarantees of free speech also come from law and policy, but in this case, by laws or policies that protect one in speaking as one wishes to (Pettit, 2018). The First Amendment to the US Constitution prevents Congress from making laws that could prohibit certain kinds of speech (Simpson, 2016), and in this sense protects it.
In theory, if a state had no laws prohibiting speech of any kind, and had a constitution guaranteeing that no laws could be created to prohibit speech of any kind, then, for the people living within that state, they would have complete and fully extensive scope to political freedom to speak without obstruction, and would have guarantee that this will not change. There would be no constraint on the extent to which the people in that state may exercise their civil rights to speak, and this civil right could not be removed. However, in practice each nation has its own laws limiting speech in various ways and to various extents, and different constitutions provide different kinds of guarantee that the scope of the right to free speech will not be constrained further. In the university sector in the UK, there is one set of laws and one set of policies that set particular limits on the scope of speech, and one policy guaranteeing free speech. Let's begin with the guarantee.
The notion of guaranteed or protected speech is evident within the Education Act 1986, section 43, which states that UK universities have a legal duty to 'take such steps as are reasonably practicable to ensure that freedom of speech within the law is secured for members, students and employees . . . and for visiting speakers'. 3 This policy is specific to the university sector, and has different implications for the different actors it mentions. For academics, it guarantees to protect academic freedom from interference from various bodies and institutions, including their own university, external pressure groups and government officials. It entitles academics to dictate, as is reasonable, the 'fundamental content of [their] research and publication' and 'the content and terms of teaching' (Moody-Adams, 2015, p. 101). For students, it guarantees that their normal civil rights to freedom of speech will be protected, and where they engage in academic research, to be granted academic freedom. For visiting speakers, it ensures that their scheduled talks will not be cancelled unfairly or unreasonably on grounds of what they will be saying.
The most well-known defence of freedom of speech, which is often attributed to John Stuart Mill (2008Mill ( [1859, is that only through free and open discussion can we promote truth and can people arrive at truth. Mill argues that since full and open discussion will promote the truth, a society should not merely tolerate speech that is objectionable, but should embrace it for the purposes of discovering what is true, and showing what is false. 4 He outlines several reasons for this. First, we should always be open to the possibility that non-mainstream views might be true, since to 'deny this is to assume our own infallibility' (p. 128). Second, even a false opinion may contain some truth, and the mainstream view may be improved by learning from mostly false opinions. And third, even if the mainstream view is wholly true, it will 'be held in the manner of a prejudice, with little comprehension or feeling of its rational grounds' unless it is 'vigorously and earnestly contested' (p. 128). Since the university occupies a particular place in the public sphere as a domain for the sharing and debating of ideas and opinions, then on the grounds of Mill's argument, the special protection offered by section 42 of the Education Act seems justified and necessary. 5 Two important issues must be noted concerning this duty imposed by the Education Act. First, it is on these grounds that one might object to activities such as noplatforming, since these involve denying certain people the opportunity to speak at universities, and such activities may be in tension with this legal duty. Second, the duty only extends to freedom of speech that is 'within the law'. And this introduces the first limit on the scope of freedom of speech in the UK, which has particular laws concerning what kinds of speech are illegal. Hence, though there are laws protecting free speech, there are limits where legal hindrances already exist. The particular laws at issue here concern those identified within the Equality Act 2010. 6 This act prohibits discrimination against, and harassment and victimisation of, individuals on grounds of age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race, religion or belief, sex and sexual orientation. Speech that falls within these prohibitions is not required to be protected by any legal duty upon UK universities. Moreover, certain instances of no-platforming may in fact be compatible with the legal prohibitions of the Equality Act, on the condition that what the speaker will say is discriminatory, harassing or victimising of one of the identified social categories. Universities must balance their duty to protect speech under the Education Act and disallow certain speech identified within the Equality Act.
The second limit on the scope of freedom of speech is given by the Counter-Terrorism and Security Act 2015. Under section 26 of the act, there is a 'Prevent duty' on universities to 'have due regard to the need to prevent people from being drawn into terrorism'. 7 Specific advice has been given to universities in the UK for carrying out this duty. For higher education institutions in England and Wales, when these institutions are considering whether or not to invite a particular speaker to give a talk, the government recommends that they 'should consider carefully whether the views being expressed, or likely to be expressed, constitute extremist views that risk drawing people into terrorism or are shared by terrorist groups'. The government also advises that a similar consideration should be taken by student unions and societies. It is advised that universities should put in place policies that 'set out what is expected from the student unions and societies in relation to Prevent including making clear the need to challenge extremist ideas which risk drawing people into terrorism'. 8 These steps to carrying out the Prevent duty for universities may seem reasonable but give rise to two problems. The first concerns the definition of 'extremism'. In its guidance for schools and childcare providers, extremism is defined as 'vocal or active opposition to fundamental British values, including democracy, the rule of law, individual liberty and mutual respect and tolerance of different faiths and beliefs'. 9 This Silencing and freedom of speech in UK higher education 5 definition is problematic in several respects, but one issue is that the definition is ambiguous and could be interpreted so as to encompass many views and ideas not usually concerned with drawing people into terrorism. For instance, recent political philosophy has seen some challenges to democracy. Jason Brennan (2016) and Alex Guerrero (2014) have both argued for alternatives to universal suffrage. On one interpretation of the guidance given by Prevent, universities should consider carefully whether or not Brennan or Guerrero should be allowed to speak on these issues, but their work is not aimed at, nor is it reasonable to interpret it as aiming at, drawing people into terrorism.
A second issue is the inadvertent effect that the Prevent duty has had on students in universities. These effects were reported to the JCHR during their investigation in 2018. One reported issue is that 'students are dissuaded from setting up events . . . out of fear of being referred under Prevent for mistakenly inviting "extremist" speakers'. A second effect has been that Muslim students 'have been dissuaded from becoming involved in student activism out of fear of being reported under the Prevent duty for expressing opinions on certain issues' ( §70-72). Implementing the Prevent duty in university policy has hindered students in speaking as freely as they would otherwise wish to.
This second issue complicates the debate in an important way. There may be no legal prohibition under either the Equality Act or the Counter-Terrorism and Security Act from saying particular things or being involved in certain societies. Nevertheless, some students are afraid of saying these things or being involved in these societies for fear of being misidentified as extremist. It may also be that students are fearful of being interpreted as in contravention of the Equality Act if their views appear discriminatory. There is an important tension here, for the students will often be free from any legal prohibition to express their views. And yet their speech is constrained through fears of the consequences of identifying oneself with views deemed to be controversial. This problem falls outside the scope of laws or policies constraining speech, but is nonetheless highly relevant to the current debate in higher education. Freedom of speech, as I have construed it, is a legal right that is protected in most cases. But since these students have free speech rights, the constraints on their speech cannot be understood as restrictions to speech in the form of law. One way of interpreting them, which is helpful for distinguishing this and other cases, is as instances of silencing: if one is not breaking a law in sharing certain views but a policy context creates fears of sharing these personal views, then that context silences these voices. In the next section, the taxonomy I offer for categorising different kinds of silencing will make sense of these and other problematic issues.
Before moving on, we need to highlight a useful distinction between restrictions to free speech and violations of free speech. The distinction between the two is one of justification. Laws and policies restrict our speech, but only violate our right to free speech if they are unjustified. For instance, the Equality Act restricts speech, but is generally thought to be justified, and hence does not violate free speech. Free speech is violated when law and policy restricts or constrains speech, but in a way that is unjustified. One way to understand this is in terms of Mill's famous 'harm principle': The only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant. (Mill, 2008, p. 94) Supposing acts restricted by the Equality Act-discrimination, harassment and victimisation on grounds of protected characteristics-would constitute kinds of harm. Then, power is rightfully used to restrict these acts. But if these are not kinds of harm, then power is not rightfully used to restrict them and hence violates the right to perform these acts. There is a way of interpreting Mill such that he would see the restrictions imposed by the Equality Act as rightful. On a popular reading of Mill's understanding of 'harm', we do harm to someone when violating their 'rights-based interests' (Thomas, 1983). Quite plausibly, people have a right not to be discriminated against, harassed or victimised on grounds of protected characteristics, and hence such acts would constitute harm for Mill, and would be justifiably restricted on his principle. Now, this is not the place to determine whether or not Mill would agree that these restrictions are justified. I merely want to show how the distinction between speech restrictions and speech violations will be useful for the following discussion. As we will see, some instances of silencing merely restrict speech whilst others violate our free speech rights, but there are other cases that are still relevant to this debate which neither restrict nor violate free speech.

Silencing in the UK university
Much of the recent literature on silencing construes it in a very particular way, as a form of harm to speech. For instance, it is claimed that silencing is 'to have one's communications or communicative capacities undermined or disabled in a way that is harmful' (McGowan, 2018, p. 138), where this harm can amount to an injustice that is epistemic (Fricker, 2007) or discursive (Kukla, 2014). In this way, 'silencing' is used as a term of art to denote a form of speech-related harm. In this section I will explain these accounts of silencing. However, I want to develop a broader account of silencing, one in which silencing is simply to prevent others from performing acts of speech. Construed this way, silencing someone is not necessarily a harm, although as we'll see, it often can be.
Following Langton (1993), we can taxonomise kinds of silencing in terms of Austin's (1975) distinction between locutionary and illocutionary acts (see Jacobson, 1995;Bird, 2002;and Maitra, 2009 for criticism). A locutionary act is the act of saying something meaningful in a language, as when someone says 'the time is 6.30am'. An illocutionary act is a speech act you perform in a certain conversational context, used to do different things with the words spoken (Searle, 1976). For instance, assertions put forward the propositional content of a sentence as being true. One can assert that 'the time is 6.30am' in order to tell others what the time of day is. Commands attempt to get others to do something. Someone might say that 'the time is 6.30am' in order to get someone else out of bed. And promises commit the speaker to future action. If someone says 'I will be home by 6.30am', then she has committed to arriving home by the stated time.
Silencing and freedom of speech in UK higher education 7 The following taxonomy first distinguishes kinds of locutionary silencing, before assessing illocutionary kinds.

Active locutionary silencing
Someone can be prevented from performing locutionary acts in ways that are either active or passive. We can identify at least three broad categories of active locutionary silencing. First, there is physical silencing, as when someone is gagged or imprisoned to prevent their speaking. There have been instances of the physical silencing of academics and students in some countries in the last few years, most notably in Turkey where students and academics have been imprisoned. If physical silencing in this form were to occur in UK universities without legal justification, it would undoubtedly be a free speech violation.
Second, there is pre-emptive silencing. This can happen in a number of ways. If a person, or group of people, are excluded from being able to act as informants on a particular matter due to unjust identity prejudice, then their voices are pre-emptively silenced (Fricker, 2007, p. 130). For instance, excluding women from political discourse, either by denying them suffrage or the opportunity to run for political office. Although I am unaware of any surveys and empirical studies on this issue, it could be the case that pre-emptive silencing occurs in a similar way in the university sector. For instance, suppose that when designing a university policy on equality and diversity many social groups within the university are consulted on their views of the policy. However, certain views are not solicited since they are deemed irrelevant, but on the basis of an unwarranted identity prejudice. This might be the case if, say, the views of particular working-class students are disregarded since those students are known to hold to more traditional, conservative beliefs and values, thought to be outdated and irrational. Here a potentially unjust social prejudice works against the working-class students, and in virtue of this, their opinions are not sought on a policy affecting them, and their voices are hence pre-emptively silenced. This is not a constraint on free speech, even if the university is itself responsible for the silencing, since it is not required under the Education Act to promote or facilitate freedom of speech, only to secure it, and this can be done consistently with this kind of prejudicial preemptive silencing. Hence, we have a form of silencing that is a concern for the wider issue of speech in higher education, but one that is neither a justified restriction to, nor a violation of, free speech.
There is also pre-emptive silencing that occurs regardless of identity prejudice, as when someone is disinvited to speak at an event because what they would have said is somehow unwelcome. No-platforming is a key example here. Some speakers have been invited to give talks at a university, only to be subsequently disinvited. This is a special case of a restriction to academic freedom. Although the academic voices are silenced, it may be compatible with the duty on universities to secure free speech, provided it is also consistent with the requirements of both the Equality Act and the Prevent duty. Take, for instance, a policy on no-platforming at the University of Leeds: the University [of Leeds] would not seek to prevent or inhibit spoken or written criticism of the state of Israel; it would not however allow criticism of Israel to be expressed in a 8 F. Malcolm form which was or might reasonably be taken to be anti-Semitic, just as it would not allow, to take another example, the expression of views intended to stir up religious hatred against Muslims. 10 Presumably, the University of Leeds does not prohibit speakers from making such criticism against the state of Israel because it is not deemed to be in contravention of prohibitions set out in the Equality Act. Along similar lines, anti-Semitic language is prohibited because it is in contravention of the Equality Act. Now, the latter involves a form of silencing-a silencing constraint on free speech-but one that is justified and hence not free speech-violating. However, if no-platforming occurs when it is not in contravention of the Equality Act or the Prevent duty, then it is a free speech-violating form of pre-emptive silencing. This is because there are no legal or policy grounds to justify a prohibition on the speech in question, and the university is failing in its duty under the Education Act to secure free speech for visiting speakers.
Another form of pre-emptive silencing concerns the student syllabus. If students feel that they are not the kind of person who should be taking a certain subject, then they may be disposed to remain quiet in class. This can occur if, for instance, a female student taking a subject only reads male authors and concludes that this subject is not for women. Such misperceptions are the cause of stereotype threat (Steele, 2010), and provide a compelling reason to have greater diversity within the syllabus. Again, we have a form of silencing that is not a restriction to free speech, but is nonetheless an important speech matter for universities.
A third kind of active locutionary silencing is inhibitive silencing. Rather than physically preventing speech, inhibitive silencing occurs when someone, or some group, are prevented from speaking through threats and intimidation (Langton, 1993, p. 315), or a sense of futility because they know they will be shouted down (Tanesini, 2016, p. 87). The JCHR report heard evidence of a number of speakers being intimidated through student no-platforming activities. In 2018 there were protests at a talk given by the Conservative MP Jacob Rees-Mogg, which ultimately prevented him from speaking. According to the JCHR report, 'freedom of expression is unduly interfered with when protests become so disruptive that they prevent the speakers from speaking or intimidate those attending' ( §4.42). Actions of inhibitive silencing that 'unduly interfere' with free expression are free speech violations.

Passive locutionary silencing
People can also be prevented from performing locutions in ways that are passive. There are two clear ways that this can occur. The first-prescriptive silencing through laws and policies-is familiar in discussions of free speech. Prescriptive silencing is essentially the work of the passive power of law and policy in preventing speech (Fricker, 2007, pp. 9-10). In one sense, anyone in the UK is entirely free to say whatever they are able and would like to. And yet, in another sense, no one is entirely free to say whatever they would like to since, if they do so, they may be subject to punishment. The latter sense is the passive power of law and policy. If someone directs racist epithets at someone else in a public space, they are liable to punitive action from the state. They are free to say the racist abuse they want to provided they have the ability Silencing and freedom of speech in UK higher education 9 to do so, but they are also passively silenced through law, and the criminal justice system that threatens significant punitive action if the racist abuse is spoken.
The legal restrictions to liberty imposed by prescriptive silencing have already been set out, but there may also be local university policies that prevent certain forms of speech due to their own interpretation of those laws. For instance, one university may make it an act of discrimination to ask, say, a man who has transitioned gender to not use the female toilets. In this case, local university policy on equality and diversity prescriptively silences voices deemed to be discriminatory, but in a passive, enforceable way. Unless the policies or laws that prescriptively silence are in violation of protective free speech duties, they are not free speech-violating even though they constrain speech. This issue has become contentious, however, in the case of the Prevent duty. It can appear that the laws concerning speech in UK universities are contradictory (Perfect et al., 2019a, p. 102). As detailed earlier, universities have a duty to protect free speech within the law, and another duty to constrain speech that could draw people into terrorism. On its face, there is no contradiction since the latter requirement simply sets a boundary for the former. It has appeared contradictory, however, where the Prevent duty leads people to restrict their speech in ways not intended, and not warranted, by the duty. This is the second form of passive locutionary silencing: a passive form of inhibitive silencing.
Passive inhibitive silencing was the main concern highlighted in the JCHR report. As already noted, some students censor their own speech, restrict invitations to certain speakers, and are cautious over which societies they belong to, in part due to a fear of being misidentified as extremists. Evidence given to the JCHR by Scott-Baumann, and detailed in her qualitative interviews on the nature of Islam on campus (Scott-Baumann & Perfect, 2020), states that the Prevent duty 'chills' freedom of speech. Another way we could put this claim is that Prevent silences speech, but in a passive, inhibitive way that goes beyond its remit of preventing terrorism. The aim of Prevent is to stop terrorist radicalisation, but many Muslim students are fearful of identifying themselves as Muslims. They are concerned with being misidentified, but also of being misunderstood due to widespread prejudicial views about Muslims. The concern is that there is institutionalised prejudicial judgement against certain groups that inhibits their speech, causing them to remain quiet rather than identify themselves with a particular social identity of a 'controversial' belief and value system. This institutionalised judgement is not simply aimed at Muslims, but can apply to many other groups as well. We noted the working-class student with more conservative values. Other religious groups can also be self-censoring for fear of social judgement, as can students holding to more conservative political views and opinions. Some of these fears are stoked by policies like Prevent, but in general it may be that the tendency of universities to be liberal-leaning institutions make them a challenging place for conservative opinions. This often goes beyond mere disagreement and political correctness (McGowan, 2018). Public culture can come with harsh socially punitive measures where people can be outcasts and branded as intolerant and, in some cases, as extremist. The fear of this occurring has a silencing effect that can be harmful. And if this becomes institutionalised, as Muslim students have claimed it has in their case, then we have harmful institutionalised silencing.
There is a problematic connection between the two forms of passive silencing, that makes sense of the apparent contradictoriness of the university's legal duties. Prescriptive silencing through law and policy constrains and restricts freedom of speech but, assuming it is justified, does not violate free speech rights. However, these prescriptions have caused limitations to speech that would not have been illegal, nor would they break any policy. They have inadvertently brought about inhibitive silencing, causing cultures of suspicion where students self-censor their speech due to fears of being misidentified and misunderstood. This has plausibly added to an existing culture of prejudice against conservative views and opinions. And in this sense, the university's legal duties are somewhat contradictory. They must secure freedom of speech, but their policies inadvertently create cultures where students, and possibly academics too, don't feel able to speak freely. That is, prescriptive silencing within the law causes inhibitive silencing that harms people's right to freedom of expression.
This connection between prescriptive and inhibitive silencing is troubling, since students have a right to share their views, provided they are not in contravention of any law of policy, and yet they are too afraid to enact this right. This marks a form of injustice, and suppresses the voices of some young citizens, where all young voices and perspectives are equally valuable in the public sphere (Laborde, 2017). The implications of this issue for recent reports on free speech in higher education will be explored below.

Illocutionary silencing
We now have a framework in place for interpreting instances of locutionary silencing. However, much of the recent work on silencing in feminist philosophy and philosophy of race has concerned illocutionary silencing, which occurs when a speaker is unable to carry out the speech acts that she intends to. A key example in the early debate on illocutionary silencing is sexual refusal: Sometimes "no," when spoken by a woman, does not count as the act of refusal. The hearer fails to recognize the utterance as a refusal; uptake is not secured. In saying "no" she may well intend to refuse. By saying "no" she intends to prevent sex, but she is far from doing as she intends. Since illocutionary force depends, in part, on uptake being secured, the woman fails to refuse. (Langton, 1993, p. 321) The idea here is that there is a failure on the part of the hearer to recognise the speech act, but that performing the speech act of refusal depends, in part, on the hearer's recognition. This failure creates what Langton calls an 'illocutionary disablement'-the intended illocutionary act of refusal is disabled and the refusal no longer counts as a refusal. In a sense, this disablement illocutionarily silences the speaker's intention. 11 Langton's proposal for the requirement of hearer recognition has been developed in numerous ways. For instance, Miranda Fricker (2007) and Kristie Dotson (2011) show how sometimes speakers are not properly recognised as knowers, and so their assertions are not recognised for what they are. If people are unable to perform assertions, then they are unable to share their knowledge. Drawing on the work of Patricia Hill Collins (2000), Dotson says that certain 'controlling images' stigmatise black Silencing and freedom of speech in UK higher education 11 women by shaping prejudicial perceptions around which they are perceived only as 'mammies, matriarchs, welfare mothers, and/or whores' (Dotson, 2011, p. 242). Dotson notes that 'the "controlling images" of black women . . . facilitates a recurring failure of audiences to communicatively reciprocate black women's attempts at linguistic exchanges by routinely not recognizing them as knowers' (p. 243). As with Langton's case of refusal, a failure on the part of the hearer to recognise the black women as making assertions could 'disable' that very speech act, making it not count as an assertion, meaning that the women are unable to share their knowledge with others.
If this account is correct, then failures on behalf of hearers to recognise the communicative intentions of speakers illocutionarily silences their speech. This is a matter of injustice when the failure is due to an unwarranted prejudice, and is the reason why silencing is always construed as a harm-because illocutionary silencing is typically an injustice that brings about harm. One particular problem with being illocutionarily silenced in this way is that it can create dispositions towards silence in the speakers. As McGowan explains: When speakers are aware that their assertions will not be given their due, speakers may decide against speaking. Moreover, speakers are more likely to decide against speaking when the credibility deficit is known to be extreme. In such cases, one might decide that speaking is hardly worth the effort. (McGowan, 2018, p. 141) This can be particularly harmful when, for instance, victims of rape decide not to speak out because they believe their testimony will not be given sufficient uptake, and when victims of racist abuse remain silent because, similarly, they expect their reports to be dismissed out of hand. So, not only is illocutionary silencing often a form of injustice (both discursive and epistemic), but it causes harm by, in Doston's words, 'quieting' its victims. Now, if someone is deemed to be irrational, or to belong to a social group that is deemed irrational, then it is quite possible that the person's testimony is not being recognised as testimony. She is not recognised to be performing assertions, but merely to be reporting what irrational people falsely believe. 12 Hence, her assertions may fail-she may become a victim of illocutionary disablement. This has not been the subject of recent empirical work in UK higher education, but some of the findings that give evidence of passive inhibitive silencing could indicate that illocutionary disablement does occur. Where students with 'conservative' views, including religious and working-class students, have felt inhibited in sharing these views due to prejudice, it could be that they are deemed to be inherently irrational. When sharing these views, they might not be considered to be performing acts of assertion, but simply to be reporting what irrational people believe. We can derive two problems from this. First, their speech acts would fail due to lack of audience reciprocity. And if the basis of non-receptivity is due to an unwarranted prejudice, then there is an epistemic injustice. Second, the failure of their speech acts might cause the kind of quieting that Dotson and McGowan are concerned about-the speakers may become timid in sharing their views because they believe they will be quickly dismissed.
Illocutionary silencing is not a natural issue of concern in the discussion over freedom of speech (Jacobsen, 1995, p. 76). There may be no unjustified constraints imposed on anyone through illocutionary silencing. However, its 'quieting' causes are an issue for the university's duty to secure freedom of speech for its members and visitors. Students should not feel as though they shouldn't speak because their views will not be given proper credence. Hence, though illocutionary silencing is overlooked by empirical work on freedom of speech, it is relevant to understanding wider issues facing unnecessary obstacles to this right. Below, I will explore one way of removing these obstacles.
We now have a taxonomy within which to examine the findings from some recent surveys and interviews, highlighted previously, and to assess the extent of the problems concerning freedom of speech in UK universities. This can be mapped as follows:

Locutionary silencing
Illocutionary silencing (disablement-quieting) Active Passive Physical Pre-emptive Inhibitive Prescriptive Inhibitive We will apply this framework to problems of speech in UK universities.

Reassessing the challenge
We have seen that some quantitative surveys revealed concern that free speech is 'under threat' in higher education. This concern is rejected as overinflated by the JCHR report from 2018, and a subsequent report from the thinktank Theos, who both maintain that media reportage has exaggerated the issue. So, is or isn't freedom of speech unduly suppressed within higher education? The answer to this question depends crucially on the ways in which speech is constrained, and we can now usefully categorise these constraints within the silencing framework.
To begin with locutionary silencing, it seems that there are very minimal restrictions to freedom of speech from physical silencing in the UK context, and hence this issue should be of little concern.
Some cases of pre-emptive silencing also appear to be negligible in number, and where they do occur, they are justified. Where pre-emptive silencing concerns activities such as no-platforming, only a few notable instances have been reported. Whether a speaker is no-platformed is a decision usually taken by the National Union of Students (NUS) in light of their policy on the issue. Of the no-platforming policy, the NUS says that it prevents individuals or groups known to hold racist or fascist views from speaking at NUS events. It also ensures that NUS officers will not share a public platform with individuals or groups known to hold racist or fascist views. 13 Most would agree that this policy provides a justified constraint on free speech, and is not in tension with the university's duty to secure freedom of speech. But despite the considerable media reportage on no-platforming, the NUS do not appear to have denied any speakers a platform to speak that goes beyond this policy. Even though some high-profile cases, such as Germaine Greer and Peter Tatchell, have made headlines, they were both able to give their talks.
Silencing and freedom of speech in UK higher education 13 The confusion appears to be between formal no-platforming activities and studentled disruption at some talks, as with the case of Rees-Mogg. However, this confusion is between pre-emptive silencing, within which no-platforming belongs, and active inhibitive silencing, which is a different category altogether. Given the lack of free speech-violating instances of no-platforming, the JCHR appear warranted in their claim that, with respect to no-platforming, 'the narrative that "censorious students" have created a "free speech crisis" in universities has been exaggerated'. So, freedom of speech is not unduly suppressed within higher education with respect to pre-emptive silencing in the form of no-platforming.
One of the other forms of pre-emptive silencing is harder to measure and, indeed, has not been the subject of empirical investigation. This is the issue of only selecting representatives from certain groups to formulate policy affecting all students and university staff. This issue is historically problematic. Women in particular have not been properly represented in public matters, and have often been intentionally excluded. Even though this issue does not concern constraints or violations of free speech, it is nevertheless a pertinent issue for the wider concern over the representation of different voices and perspectives in the university. It therefore needs to be highlighted as a potential issue of pre-emptive silencing that requires serious consideration and further empirical research.
A third form of pre-emptive silencing is more widely accepted-when non-inclusive university syllabi produce stereotype threat. The move towards greater diversity and inclusivity within the syllabus-connected to the notion of 'de-colonising' the curriculum-is a known issue, and one that can create pre-emptive silencing.
The final form of active locutionary silencing is inhibitive silencing, which occurs when people are shouted down, threatened or intimidated from speaking. Here we have a kind of silencing that, if it does occur, would be free speech-violating. But again, it is difficult to measure how widespread this issue can be. Amongst students in classrooms, it is the role of the lecturer to facilitate respectful debate, and the duty on universities to put in place complaints procedures if students feel victimised or harassed.
Amongst talks given by visiting speakers, there have been a small handful of cases where the speaker has not been able to deliver the talk due to being shouted down. In addition to Rees-Mogg, Maryam Namazie had her talk disrupted by heckling students at Goldsmiths, University of London. Although these are both cases of active inhibitive silencing that are free speech-violating, they are extremely isolated cases. So, freedom of speech with respect to physical inhibitive silencing is, for the vast majority of visiting talks, not unduly suppressed within higher education.
The more problematic issue concerns passive inhibitive silencing. As with wider society in the UK, more traditional views on issues including marriage and gender are being viewed as discriminatory. Such views are often linked to class and religious identity, and hence there is more stigma around identifying oneself with these views in public places, including the university. Perhaps it is for this reason that, as the research mentioned above shows, the majority of undergraduate students do not feel completely free to express their views and opinions on campus; that many people believe freedom of speech to be 'under threat' and that a majority of Muslim interviewees self-censor their own speech. But we must carefully distinguish two forms of passive silencing here. The first-prescriptive silencing-occurs when policy and law prohibits speech. If the views being expressed are discriminatory, in line with the Equality Act, then they are prescriptively silenced, not inhibited. The distinction is artificial, but important on grounds of justification. If justified law and policy restricts speech, then it is not free speech-violating, but if people are unable to share their views despite these views not infringing on law or policy, then we have a free speech violation.
The issue, then, is the extent to which people are unable to freely express their views and opinions where doing so does not contravene any justified law or policy. Quite plausibly, this is a genuine concern. As Scott-Baumann and Perfect (2020) found in their interviews, some students do not even want to identify as Muslims for fear of prejudice and social sanction. This is partly due to activities stemming from the Prevent duty, as highlighted in the JCHR report: some of those giving evidence to us said that the Prevent duty guidance for higher education institutions inhibits free speech. The fear of being reported for organising or attending an event, combined with the increased levels of bureaucracy following the introduction of the Prevent duty, is reported to be having a "chilling effect" on freedom of speech. The Committee acknowledges the need for a strategy to prevent the development of terrorism both in universities and in wider society, however, we repeat our call for an independent review of the Prevent duty. This review should include consideration of its impact on free speech in universities particularly on Muslim students but also on students of other faiths or no religious faiths. (p. 5) My proposal is that we understand this 'chilling effect' in terms of passive inhibitive silencing. Students being unwilling to organise events, or even to self-identify as Muslim, or generally religious, or to associate oneself with views deemed to be controversial is a genuine concern. This often does not contravene law or policy and is hence free speech-violating.
We have so far been able to answer whether freedom of speech is unduly suppressed within higher education by working with distinct categories of silencing. There is little concern with regard to the forms of physical silencing, but passive inhibitive silencing does appear to be problematic. These distinctions explain the contradictoriness of the research outcomes highlighted earlier. Freedom of speech is not 'under threat'-it is not unduly suppressed-in terms of active silencing, including forms of no-platforming. However, research indicates that it is unduly suppressed in terms of passive inhibitive silencing. These distinctions are alluded to, but not formally made, in the report by Theos. Their conclusions are that: Freedom of speech on campus is not in crisis, but a minority of students feel under pressure to self-censor their views. The crisis narrative about freedom of speech is largely exaggerated. Most students feel free to express their views. However, there are some factors which can chill freedom of speech on campus, including student activities (like intimidating behaviour of protesters) and structures like the Prevent Duty. Some students feel under pressure to self-censor their views, including students with socially (or politically) conservative views; students who support the policies of the State of Israel; and some (though not all) Muslim students who feel unfairly targeted by the Prevent Duty. (Perfect et al., 2019b, p. 3) Silencing and freedom of speech in UK higher education 15 These conclusions are made clearer by the silencing distinctions. However, two criticisms are forthcoming. First, although the students who feel under pressure to self-censor their views may be in the minority, this minority may be larger than is implied here. It may encompass many or most religious students, and students from certain class or cultural backgrounds. Second, there are concerns relevant to the representation of different voices on campus that, though constrained, are not free speech-violating, and are hence ignored by this report. As noted in the foregoing, one possible issue concerns pre-emptive silencing in the form of the curriculum. Another problem concerns testimonial quieting. This latter concern stems from illocutionary silencing, which is the final issue to address.
The extent of illocutionary silencing may be impossible to empirically track. What might be easier, however, is the extent of testimonial quieting and, in particular, some of the causes of it. Testimonial quieting occurs when speakers decide not to speak because they believe their intentions to communicate knowledge will be misunderstood, and their speech acts will be disabled. A stark example of this quieting is the disvaluing of the testimony of victims of rape and sexual abuse. Cases like this typically occur due to harmful prejudices that work against the testifier. In the case of those testifying about being victims of sexual assault, there is often a prejudicial stereotype of such people as attention-seeking or na€ ıve, which can cause hearers to be disposed towards discrediting their testimony. This in turn leads some victims to keep their experiences quiet, due to an expectation that their testimony would not receive sufficient uptake or be given impartial credence.
Within a university context, some factors may also lead to harmful stereotypes that produce credibility deficits. One factor that has been highlighted recently by sociologist Adam Dinham is a failure of religious literacy: I have observed a lamentable quality of conversation about religion: at the same time, a pressing need for a better quality of conversation in order to avoid knee-jerk reactions which focus only on "bad" religion. (Dinham, 2016, p. 205) A failure to understand religions properly can lead to stereotypes that characterise religious people as inherently irrational, illiberal, inegalitarian or extremist. Someone holding this kind of stereotype may well be disposed to have a credibility deficit towards the testimony of religious people, which may be unwarranted. Religious people who then experience continual rejection of their testimony may become testimonially quieted and insulated within religious groups.
Improving religious literacy may help to overcome unwarranted stereotypes that create these kinds of credibility deficits, and in turn create cultures where religious people feel more able to express their own views and opinions without facing disregard out of hand. This literacy could be part of the wider attempt by universities to break down harmful stereotypes that lead to implicit bias. Challenging sexist, racist and homophobic biases is critical for overcoming unwarranted prejudices, which can lead to epistemic injustice and subsequently cause testimonial quieting. And yet harmful biases also affect students from religious and working-class backgrounds. One way to overcome some of these biases is through proper literacy on these matters, and through diversification of the syllabus.

Conclusion
The current issue of free speech in higher education can be fruitfully explored through the notion of silencing. Within a fully developed silencing framework, we can see where free speech is unduly constrained and where it is not. This resolves the apparent tensions across recent qualitative interviews and quantitative survey data. Passive inhibitive silencing, partly caused by prescriptive laws and policies, creates climates of self-silencing that is a genuine concern, and requires attention from experts in public policy. But conceptual work on silencing also shows us where there have been speech matters in higher education that are not violations of free speech. Preemptive silencing in the form of proper representation on policy matters, and in creating diverse syllabi, are both critical issues, each requiring further empirical and conceptual exploration.

Ethical guidelines
Ethics approval was not required for this research.

Conflict of interest
The author declares no conflict of interest.

Data availability statement
Data sharing is not applicable to this article as no new data were created or analysed in this study. NOTES acknowledge the act and understand the speaker's intentions. This view is shared in other work in addition to Austin. Jurgen Habermas has argued that a speech act only succeeds in reaching understanding for a hearer when she takes up 'an affirmative position' toward the speech act (Habermas, (1984, pp. 95-97). 12 Fricker calls this 'epistemic objectification' (Fricker, 2007, p. 133). 13 See NUS' No Platform Policy: Key information, February 2017. They list six no-platformed organisations: al-Muhajiroun, British National Party (BNP), English Defence League (EDL), Hizb-ut-Tahir, Muslim Public Affairs Committee and National Action.