Truth, Discussion, and Free Speech in On Liberty II

Abstract

In this article, I offer a reading of On Liberty II which focuses on the structural features of the argument that Mill presents. Mill's argument, I suggest, is grounded on an appeal to the value of truth, and is divided into three sub-arguments, treating true, false and partially true opinion respectively. In section 1, I consider what constraints the teleological orientation of Mill's argument places on the case he makes, before examining in section 2 what the division of Mill's argument into three exhaustive sub-arguments tells us about the nature of ‘discussion’ as Mill uses the term. I go on, in section 3, to suggest that although On Liberty II does not offer a defence of free speech in the broad sense in which the term is often now used, we should be optimistic about the chances of finding such a defence in On Liberty III.

The argument of On Liberty II is generally well known. The chapter, setting aside a brief introduction and conclusion, is made up of three ‘divisions’¹ and runs as follows:

For any opinion considered as a candidate for suppression,

• true: that opinion might be true, in which case it should be not be suppressed (On Liberty, XVIII: 229–43),

• false: that opinion might be false, in which case it should not be suppressed, for its airing can contribute to a better justification and understanding of the truth (On Liberty, XVIII: 243–52),

• partially-true: that opinion might be partially-true, in which case it should not be suppressed, for its airing can help us achieve the whole truth (On Liberty, XVIII: 252–57).

Therefore: there should be no suppression of the discussion of any opinion.

The effectiveness of the overall argument of course depends on the effectiveness of each of the sub-arguments for true, false, and partially-true – and this question has received much attention in the secondary literature.² In this article, I wish to attempt a different approach to On Liberty II. Rather than focusing on the detail of these sub-arguments, I wish here to consider certain structural features of the argument overall. In particular, I wish to examine what the form that Mill's argument takes can tell us about the scope of ‘Freedom of Discussion’ as defended in On Liberty II, and the extent to which we must rely on arguments beyond that chapter to support free-speech rights as they are often understood today.

I begin, in section 1, by highlighting the oddity – often missed – in Mill's appeal to the value of truth as the ground of the argument Mill offers in On Liberty II. Very little of detail is said in that work about how truth fits into Mill's overall scheme of values, or how it is to be balanced against the existence of other values. I suggest that the most plausible interpretation of the kind of argument that Mill gives in On Liberty II limits its aim to that of establishing norms of freedom internal to a specific kind of practice. I go on, in section 2, to consider what the form of argument deployed in On Liberty II tells us about the boundaries and nature of that practice – and the extent to which Mill's defence of freedom within that practice can support rights to free speech in the sense that the term is often used today. I conclude, in section 3, by suggesting that for a defence of free speech in this broader sense, we must look beyond On Liberty II, and that though we should be optimistic about the chances of finding such a defence in On Liberty III, that defence will be a limited one.

1. The teleological orientation of On Liberty II

The argument of On Liberty II is primarily epistemic in nature. By that, I mean the following: in arguing for liberty of discussion, Mill attempts to establish that there exists a relation between free discussion and our epistemic good, and he appeals to the value of the latter in motivating the former. Mill's conception of the epistemic good of human beings is a broad and liberal one. For convenience, I will refer to it as knowledge of the truth, but we should keep in mind that when Mill speaks of knowledge in On Liberty II, he means to indicate not merely a true belief, but a true belief held ‘the way in which truth ought to be held by a rational being’ (On Liberty, XVIII: 244). This involves the ability, Mill claims, to connect a proposition to its grounds and to its implications, to be able to argue against alternatives, and to meet counterarguments offered by others. His ideal is that of true belief held as ‘living truth’ rather than ‘dead dogma’ (On Liberty, XVIII: 243). Nevertheless, it is on the basis of its relation to something valuable – knowledge, in this demanding sense – that Mill argues that our discursive interactions should be governed by norms of free discussion.

There is, I think, a puzzle about the nature of the relation Mill aims to establish between free discussion and knowledge of the truth. Mill certainly does not believe that free discussion of an opinion is sufficient to bring about knowledge of the truth of that opinion: he is well aware that the existence of free discussion over any given length of time is quite compatible with the prevalence of false beliefs. But he does at times seem to move between the claim that freedom of discussion is a necessary condition for knowledge and the claim that freedom of discussion is positively conducive to knowledge of truth. I will not here attempt to resolve the puzzle of which he regards as the core claim.³ For the purpose of this article we might remain neutral on the particular connection Mill envisages between free discussion and knowledge of the truth by thinking of it as an enabling relationship, while noting that ‘x is enables y’ can be read either as x is conducive to, or necessary for, y. The important thing, for our purposes, is as follows: Mill makes no claim about an ‘abstract right’ to air one's opinions in On Liberty II. Nor, tellingly, does he make any claim to the effect that discussion should be free because it can cause only offence, but never harm.⁴ It is by appealing to a particular end that he regards as valuable, and showing which conditions enable this end to be achieved, that he means to convince the reader that there should be absolute liberty of discussion. The argument, that is to say, is teleological in orientation.

Bringing the underlying teleological orientation of the argument into focus is useful – for if the argument is to be regarded as valid, as Mill obviously intends, then this orientation imposes significant constraints on how it can be interpreted. Clearly Mill does not think that freedom of discussion is attended by no other effect than that of enabling knowledge of the truth; neither does he hold that knowledge of the truth is the only thing of value for human beings. But a teleological argument which appeals solely to the value of knowledge to offer a vindication of freedom of discussion tout court could only be valid on exactly those grounds. If, for instance, we were to establish that experimentation on humans stood in an enabling relationship to knowledge of the truth, we could not thereby seamlessly infer that experimentation on humans should be permitted. This would only be a reasonable conclusion to draw if we were confident that human experimentation had no other effects beyond those of enabling knowledge, and that were no other values which deserved consideration.

Mill confines his attention in On Liberty II to the consequences of freedom of discussion for human beings considered from a distinctly epistemic perspective, arguing for this liberty on the basis of our epistemic good as human beings. He was, however, well aware of the dangers of drawing all-things-considered normative conclusions on the basis of overly narrow abstractions – he is critical, throughout his work, of those who attempt to deduce whole truths from partial premises.⁵ It is unlikely that he made this mistake. How, then, should we read his argument?

In order to understand how Mill's argument could be considered valid, it will be useful, first of all, to note his understanding of the structure of human activity as outlined in System of Logic VI. Here Mill claims that each human practice – ‘art’, in his terminology – is properly governed by a body of rules which specify how to achieve the end at which that practice aims.

Every art has one first principle, or general major premise, not borrowed from science; that which enunciates the object aimed at, and affirms it to be a desirable object [. . .] The art proposes to itself an end to be attained, defines the end, and hands it over to the science. The science receives it, considers it as a phenomenon or effect to be studied, and having investigated its causes and conditions, sends it back to art with a theorem of the combination of circumstances by which it could be produced. (System, VIII: 949, 944)

Each art, then, is formulated as a body of rules offering guidance on how to bring about, or sustain, a specified desirable end. Architecture, for instance, is a body of rules aimed at bringing about ‘beautiful or imposing’ buildings; medicine is a body of rules aimed at effecting the ‘preservation of health’ (System, VIII: 949). We might term the ends of the individual arts proximate ends. There are, Mill suggests, many valuable proximate ends in life.⁶

The rules of action specified by each art are local to that practice. Arts offer guidance on how to act if one wishes to bring about the specified end – they are, in this sense, ‘hypothetical imperatives’, to draw on Kantian terminology. But proximate ends, and therefore the rules of individual arts, will often come into conflict. It is the job of the grand commanding ‘Art of Life’ – ‘Practical Reason’, as Mill also calls it – to determine the ‘place in the scale of [these] desirable things’ overall, and to reconcile conflicts when these ends clash. This determination, Mill suggests, should be made on utilitarian grounds. ‘I do not mean to assert that the promotion of happiness should be itself the end of all actions, or even of all rules of action. It is the justification, and ought to be the controller, of all ends, but is not itself the sole end’ (System, VIII: 952).⁷

A picture therefore emerges of many and various human practices, each with distinctive ends, guided by rules which help achieve those ends and underpinned by their place in securing human happiness. The question posed above was this: how can we read Mill's argument as valid, given that a tout court vindication of freedom of discussion would involve considerations of all effects and ends, but that Mill considers only epistemic effects and ends? The answer, I suggest, is that Mill did not mean to offer a tout court vindication of freedom of discussion. Rather, he offers an account of the rules we should adopt as effective for discussion, given the aim of that practice. Mill focuses his attention in On Liberty II only on the effect relevant to one valuable end, because discussion – in the sense that Mill uses the term – is only one human practice among many.

We should certainly value knowledge of truth, and pursue it as an end. As is the case with other things which we value, however, sometimes this end will conflict with other valuable ends – and as such the rule against informal or legal suppression of discussion will be incompatible with some other rule facilitating another end. (Such is, in effect, a more generalised way of framing the classic worry that the free discussion of some views might endanger security, breach copyright, violate privacy, undermine social cohesion, etc. – and that our commitment to free discussion might therefore clash with our other commitments.) In such a case, the job of determining the priority of our ends, as noted, falls to the ‘Art of Life’: we are called upon to find a balance between valuable proximate ends on the basis of their contribution to the overarching end of human happiness.⁸

This is not to make the claim that freedom of discussion should often be disregarded on the grounds of its incompatibility with other valuable ends. Indeed, it may be that, such is its value, knowledge of truth should almost never be sacrificed in favour of some other end. Clearly, Mill thinks the end extremely important, and takes the injunction against suppression of free discussion extremely seriously:

I deny the right of the people to exercise such coercion, either by themselves or by their government. The power itself is illegitimate. [. . .] If all mankind minus one, were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind. (On Liberty, XVIII: 229)

But it is equally clear that he does not think that this injunction should be exceptionless – for he holds that no exceptionless practical rules are possible. ‘[R]ules of conduct cannot be so framed as to require no exceptions’ (Utilitarianism, X: 225). ‘[T]he admission of exceptions to rules is a necessity felt in all systems of morality’ (Whewell, X: 182). ‘To admit the balance of consequences as a test of right and wrong, necessarily implies the possibility of exceptions to any derivative rule of morality which may be deduced from that test’ (Taylor's Stateman, XIX: 638).⁹ A commitment to knowledge as an important object of value does not mean that in each and every instance knowledge should take precedence over other objects of value.

The question of how and when to balance knowledge against other proximate ends must be answered not summarily and a priori, but with serious consideration to the role that various domains of knowledge of truth play in securing human happiness. Mill simply does not attempt to treat these sorts of issues in On Liberty II – rather, he confines his discussion only to the internal rules which enable discussion to lead to truth. He notes in On Liberty that the ‘truth of an opinion is part of its utility’ (XVIII: 233) – a view echoed elsewhere¹⁰ – but offers little by way of expansion, explanation, or defence of that claim. Because he never offers a sustained treatment of the issue, it is hard to specify precisely where Mill thinks that knowledge of truth falls in the ‘scale of desirable things’, and therefore hard to determine how, in his view, we should approach cases in which free discussion conflicts with other ends. But it seems clear, because of the teleological structure of the argument of On Liberty II, that we should be open to the fact that such judgements will occasionally have to be made.

2. The nature of discussion

I have suggested that, in arguing for Freedom of Discussion, Mill offers an account of the rules we should adopt within a practice, given the end specified as valuable by that practice. In this sense, ‘Discussion’ names a practice – a certain kind of activity – and is a technical term for Mill.¹¹ We might ask, therefore, what the boundaries of discussion are. What counts as discussion, in this technical sense?

It is clear that Mill regards discussion as a cognitive activity: an activity, as we have noted, that has knowledge of truth as its fundamental aim. Mill argues that consideration of opinion – whether true, false, or partially-true opinion – stands in an enabling relationship to knowledge, and that all contributions to discussion should therefore be permitted. The move from true, false, or partially-true opinion, to all contributions to discussion is telling. It implies that Mill regards the consideration of true, false, or partially-true opinion as exhaustive of discussion, for otherwise the inference from these three cases could hardly be secure. Mill's argument is that we should be free to engage in the discussion of an opinion because discussion enables us to achieve knowledge of the truth regarding that opinion. But this suggests that discussion involves consideration of claims which are able to be true. Discussion, that is to say, is constituted by the statement and consideration of truth-apt propositions.

So framed, the promise of ‘freedom of discussion’ seems extremely limited – narrow, certainly, by comparison to the breadth of activity that is defended under the mantle of ‘freedom of speech’ in public discourse and in law. Consider, for example, that in U.S. First Amendment law the burning of a flag is regarded as protected speech.¹² Certainly, the burning of a flag may be occasioned by the consideration of some truth-apt proposition about the world. But it makes little sense to think of burning the flag itself either as an act of consideration or as the sort of thing to which a truth-value can be assigned. The action may be evaluated as appropriate, inappropriate, just, tasteless, or patriotic, but it clearly does not state a truth-apt proposition – it does not make a claim, but rather signals protest, anger, and perhaps a call to resistance. Nor is it the public consideration of some claim. The activity is primarily an expressive rather than a cognitive one, and so is not covered by the argument of On Liberty II.

Further examples of the broadness of the coverage of First Amendment law by comparison to that which can be grounded on Mill's argument are easy to replicate. Under Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, the Supreme Court ruled that marching in a parade qualified as protected speech. Marching in a parade, as the Court acknowledges, itself conveys no ‘succinctly articulable message’, but is rather an instance of symbolic expression, and protected as such under the category of speech. So, too, other purely expressive acts, in the form of visual art and instrumental music, receive protection under the First Amendment. The ‘painting of Jackson Pollock’ and the ‘music of Arnold Schoenberg’ are regarded as ‘unquestionably shielded’ as speech in Hurley despite the fact that they are incapable of conveying a ‘particularized message’.¹³ Such expressive works cannot be construed as the consideration of opinion which is capable of truth, falsity, or partial truth, as per the categories that form the basis of argument in On Liberty II.

But although Mill's argument is limited to protecting discussion as the statement and consideration of truth-apt propositions, there are at least two ways in which that notion can be unfurled so as to cover more territory than we might initially expect. The first is simply to note that the range of statements which Mill regards as amenable to evaluation in terms of truth and falsity is a broad and ecumenical one. The cognitive is the domain of statements which can serve as evidence in an argument and which can themselves be evidenced. But, to Mill's mind, this category is not confined, for instance, only to utterances which can be translated without remainder into statements which can be directly empirically verified. Mill is clear, after all, that normative claims – statements about how there is reason to believe and act – are amenable to evidence and argument, and evaluation in terms of truth. As such they fall within the domain of the cognitive.¹⁴ There is little reason to think that Mill would deny that aesthetic claims can, similarly, fit into a network of evidencing and being evidenced.¹⁵ All of these forms of statement would be covered by Mill's account of Freedom of Discussion.

Moreover, we might note that because a claim about the status of any given linguistic act as truth-apt is itself truth-apt and therefore subject to freedom of discussion, there is an expansionist push from within the domain of the cognitive which may result in the protection of various forms of non-cognitive utterances. One might, by instinct, doubt that some statement φ – ‘apples are better than oranges’, for instance – is the sort of statement which can bear truth. Grammatical form notwithstanding, it might be claimed, the statement is merely the expression of a preference, as such excluded from the cognitive domain, and therefore from the scope of Mill's argument. But because the claim that ‘φ is not truth-apt’, is itself truth-apt, and because this statement therefore falls within the scope of Mill's argument for freedom of discussion, we must be willing to hear evidence about φ's truth-aptitude. Such evidence will include evidence for its truth, and therefore involve full discussion of φ.¹⁶ The result is generalisable, of course, and as such anything within the territory of truth-aptitude will be covered by Mill's argument.

Secondly, and more importantly, however, we should note that though discussion in Mill's sense involves the consideration only of truth-apt statements, it does not follow that consideration itself takes place only via truth-apt statements. The assertion of truth-apt propositions is clearly the foundation on which discursive consideration takes place – without individuals expressing opinion, we would not find ourselves in the position of considering those opinions at all. But consideration, in the context of dialogue, amounts to thinking together about a truth-apt proposition, and thinking together might take us well beyond the exchange of truth-apt propositions. Trivially, for instance, it may involve the posing of questions – rhetorical or otherwise – even if questions are not themselves truth-apt. (While ‘murder is wrong’ may be true, ‘is murder wrong?’ can be neither true nor false.) But other forms of truth-inapt linguistic or non-linguistic activity, too, might constitute the consideration of statements which are themselves truth-apt. If freedom of discussion involves consideration in dialogue as well as the assertion of truth-apt propositions, protection is extended to thinking as it takes place when given concrete form and externalised in the public domain.

Historically, for instance, satire has been an important medium for individuals seeking to contribute to public consideration of political ideas.¹⁷ (Curiously, given that he himself was a target in contemporary comic periodicals, we might note that Mill himself never speaks directly to the issue of satire. Reflection on this medium, however, may nevertheless help to give an example of the diverse ways in which public consideration can take place.) Satire need not involve making claims which can be subjected to evaluation in terms of truth or falsity – rather, it most often involves portrayal, whether of real or fictional subjects, in such a way as to draw attention to the injustice or absurdity of a policy or situation. By doing so, however, it can itself be an instance of thinking through that policy or situation with others: a move in an ongoing dialogue which has revealing the truth as its aim. Such a move can be made in verse, or in prose, or indeed by way of visual art and cartoons. James Gillray's The Plumb-Pudding in Danger shows Bonaparte and Pitt at the dinner table, carving the globe as if it were a roast pudding. The cartoon is clearly not itself capable of truth or falsity – but it offers a framing of the Napoleonic proposal of peace-making which draws attention to worries about empire-building and the ambition of England and France. It shows what might otherwise be overlooked: the unsavoury principles and motives which underlie the policy, the fragility of the situation, and the potential consequences of endorsing the proposal.¹⁸ As such, it contributes to the ongoing process of thinking about the prudence of a given policy. Such cartoons, that is to say, are not merely acts of self-expression, but themselves contribute to a dialogue which has truth as its end. They constitute public acts of consideration which can make advances in the pursuit of consensus and truth, and Mill's epistemic defence of freedom of discussion is therefore applicable to them. So, also, can fiction, parable, and other forms of art make contributions to the consideration of truth-apt claims: they, too, can be attempts to publicly ‘think through’ an issue with others.

The line between acts which constitute public consideration and those which are merely a form of self-expression – those which give voice to an emotion, state-of-mind, or creative impulse – is of course an extremely difficult one to draw. Making such discriminations requires a considerable exercise of judgement: that Orwell's 1984 is a contribution to a discussion which attempts to publicly think through an issue of policy, whereas Monet's Water Lilies or Sibelius's Symphony No. 2 are not, is a substantive and contestable claim; so too is the claim that Martin Luther King's The Two Americas is a contribution to a discussion, whereas the burning of a flag is not. And, of course, we must acknowledge that many actions may be both acts of self-expression and contributions to discussion – with certain aspects of their performance protected from censure by norms of free discussion, and certain aspects governed by other norms. But that there is a meaningful distinction to be made between the discussion of a truth-apt claim and behaviour which is merely expressive is certainly a position assumed by the form which On Liberty II, with its focus on knowledge, takes.

3. Freedom of expression

I have, in the first two sections of this article, offered a reading of the argument of On Liberty II which focuses on the form that the argument takes. In doing so, I have attempted to complicate our understanding of that chapter in two ways. In the first section, I highlighted the teleological orientation of the argument. Freedom of discussion is defended because of its relation to a proximate end which is assumed by the practice of discussion – knowledge of the truth. This good, however, must be balanced against other goods, and its place in the scale of ends determined by appeal to the overarching end of life. It cannot be assumed that in all cases, the good of knowledge will trump other goods. Determining when Mill's argument offers protection of discussion will involve difficult judgments about how various domains of knowledge contribute to human happiness.

Difficult judgements are also called for in determining whether something is an act of discussion, in Mill's sense. The form of argument Mill offers to establish freedom of discussion – divided exhaustively into sub-arguments addressing the cases of true, false, or partially-true opinions – indicates that he views discussion as a distinctively cognitive activity. Mill's aim is to establish freedom for the statement and consideration of truth-apt propositions – but of course not all linguistic expression is truth-apt, let alone all expressive activity. The argument does not establish protection as broad as that assumed by First Amendment understanding of ‘speech’, which shields a diverse array of behaviour, from musical performance to flag-burning. The boundaries between expression and discussion, however, when it comes to public consideration, are difficult to draw.

All of this to say: determining which forms of activity count as discussion is no easy task, and just because an action qualifies as ‘discussion’ in Mill's sense does not mean that protection is thereby seamlessly conferred. Of course, we should not allow a recognition of the complexity of Mill's argument in On Liberty II to generate doubt in our mind as to the basic liberal orientation of that work as a whole. The argument of On Liberty II does not offer protection to linguistic or non-linguistic acts of symbolic expression as such, but this is not to say that such acts do not receive protection from the arguments given elsewhere in that work. Even if purely symbolic expression does not constitute the statement or consideration of some truth-apt proposition, it may nevertheless be shielded from interference on other grounds. In particular, On Liberty III, which argues for what we might think of as the freedom to develop and express one's own personality, can be taken to offer protection to kinds of expressive activity which are not covered by the chapter it immediately follows.

On Liberty III is a paean to authenticity, originality, and self-expression, arguing that individuals should be given free scope to live according to their own nature and impulses, as opposed to merely following social expectation or convention. Mill's argument for the freedom to develop and express one's own personality, like the argument for freedom of discussion, is teleological – he attempts to justify a rule against constraints on ‘self-development’ by appealing to the proximate end promoted by that rule (On Liberty, XVIII: 266). Where members of a community are able to choose their own ways of life, he argues, they exercise their active powers of observation, reason, discrimination, judgment, desire. Freedom to explore one's own nature therefore promotes ‘well-developed human beings’, with ‘fullness of life’, a ‘strong will’ and ‘energetic character’. This proximate end he terms ‘Individuality’ (On Liberty, XVIII: 263–67). The end is to be highly prized, Mill argues, on the grounds both of its direct and indirect contribution to the final end of human happiness.¹⁹

Where, not the persons’ own character, but the traditions or customs of other people are the rule of conduct, there is wanting one of the principal ingredients of human happiness, and quite the chief ingredient of individual and social progress. (On Liberty, XVIII: 261)

The promotion of Individuality involves leaving as much room as possible for each person to expand ‘according to the tendency of the inward forces which make it a living thing’ – while preventing the growth of each to unduly impede the growth of others.²⁰

One cannot be oneself without the ability to express oneself, and the ability to express oneself – to give an external manifestation to one's internal thoughts and feelings – is so much a part of living according to one's own personality that it must be taken as a central part of what On Liberty III protects. This is so, whether expression takes the form of the assertion of truth-apt propositions or something entirely non-cognitive in orientation. Living in accordance with one's own personality will, of course, involve articulating one's opinions. But it may also involve finding an outlet in the production of poetry, dance, music, or literature. Such artistic endeavours are among the deepest and most important ways human beings have of providing a voice to their distinctive natures. It may equally well involve finding an outlet in the clothes one wears – whether a bow-tie and tweeds or a ‘F*ck the Draft’ jacket²¹ – or expressing one's anger by burning a flag. Giving air to the spontaneous expression of internal states is integral to allowing personality to ‘grow and develop itself on all sides’ (On Liberty, XVIII: 263). Such activity is therefore protected by a wholly different argument from that offered in On Liberty II.

Like On Liberty II, however, the argument of On Liberty III involves an abstraction. Mill focuses particularly on the effects of freedom on individuals’ character, and the rules we must follow to secure one particular proximate good: that of Individuality. And for that reason again, the argument cannot provide anything other than a rule which will be subject to possible exceptions – for other effects will also follow from instances of following this rule, and the end it identifies as good may conflict with other proximate ends judged valuable. As was noted above, when such conflicts between values occur, appeal must be made to the Art of Life to judge where in the ‘scale of desirable things’ these ends sit in their respective contribution to overall happiness. Such is the case for all practical rules in Mill's philosophy.

Cognitive and non-cognitive expression alike receive protection from On Liberty III. If we desire an argument from Mill for freedom of speech in the sense which encompasses both, it is to this chapter that we must turn, rather than On Liberty II. But we should note that the level of protection offered by On Liberty III is clearly not the same as that offered by On Liberty II. Mill starts his defence of freedom to develop and express one's personality by conceding that ‘[n]o one pretends that actions should be as free as opinions’ (On Liberty, XVIII: 260). Indeed, the very fact that Mill offers a separate argument in On Liberty II for freedom to state and consider opinion suggests Mill thought merely defending discussion as an expressive act which promotes the end of self-development would miss something important. The proximate ends to which On Liberty II and On Liberty III appeal are different, and Mill regards the end appealed to in the earlier chapter – knowledge of truth – as especially important, and holding a particularly high place in the scale of valuable things.

In government, perfect freedom of discussion in all its modes – speaking, writing, and printing – in law and in fact is the first requisite of good because the first condition of popular intelligence and mental progress. All else is secondary. (Diary, XXVIII: 661)

Although all forms of expressive activity receive protection, then, the statement and consideration of truth-apt propositions has protection over-and-above other linguistic and non-linguistic forms of expression. As noted above, Mill holds that no exceptionless practical rules are possible. But the rule proposed in On Liberty II should be expected to be subject to fewer exceptions than that proposed in On Liberty III. The end that discussion serves is in Mill's view particularly important, and so more will be required to justify interference in the statement and consideration of truth-apt propositions than other kinds of expressive activity.²²