How not to do things with words: On conceptions of language and freedom of speech

1 Introduction

In the 1859 publication, On Liberty, John Stuart Mill makes a strong and convincing argument for the paramount importance of a free discussion. From this position he infers that there ought to be little, if any, legal restriction of freedom of speech. He does not explicitly mention the US Constitution, 1st Amendment, but he would no doubt be happy with the wording: “Congress shall make no law (…) abridging the freedom speech (…).” ^[1] However, among other possible objections, Mill’s conception of language is inadequate: He conceives language mainly as a way of referring to the world, which is far too simplistic. Language is much more than referring, as shown by later philosophers such as Gottlob Frege and John L. Austin. The purpose of the present paper is to consider whether such updated conceptions of language may have any bearing on the understanding of freedom of speech.

2 J.S. Mill’s threefold argument

In his famous defense of freedom of speech Mill applies a threefold argument: Firstly, his argument goes, freedom of speech is indispensable – or at least preferable to prohibition and censorship – because “the opinion which it is attempted to suppress by the authority may possibly be true” (Mill 1968: 136). Nobody is unerring, and from history it is easy to see that received ‘truths’ have been no truths at all. We therefore have to acknowledge the possibility that other opinions may be true, no matter how convinced we are that this is not the case. Secondly, even if opinions are not true, we need them lest the true opinions decline into mere dogma and superficial knowledge. Mill makes the convincing point that we constantly need to be challenged even if we hold the truth. Without such challenges, we forget the reasons for the truths, and we decay into irrationality. This implies that even if we have found the truth (which according to the first and third point is not likely to happen) we still need to be open to opposing opinions. And thirdly, opinions may often be partly true, implying that even if they are wrong in some aspects, they are also right, and we need them to improve our own, merely partial, truths. For Mill, absolute truths and absolute falsities occur rarely, if ever. Most often, opinions contain some element of truth, and also for this reason opinions should not be silenced. “Truth, in the great practical concerns of life, is so much a question of the reconciling and combining of opposites …” (Mill 1968: 164), which would not work if some opinions were silenced. Mill is here quite close to Hegel’s dialectics, and he is anticipating Peirce’s and Popper’s ideas of truth as a desirable but unobtainable goal. He is also anticipating Habermas’ notion of communicative rationality and the idea of the strange powerless power of the better argument.

Mill is not only strongly in favor of freedom of speech, he is also expressly in favor of not limiting the freedom to moderate expressions: He argues e. g. that it is impossible to set the boundaries, that the most insulting speech is in fact the convincing one and that the worst offenses in a discussion are to mispresent facts and distort the opinions of the opponents, which is, alas, what is happening all the time and not possible to do away with by means of criminal liability (Mill 1968: 169–171).This idea of an almost unlimited freedom of speech also fits the Mill’s ‘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” (Mill 1968: 129).And it does not really harm one to be insulted, let alone to hear opinions with which one disagrees.

3 The yoke of the public opinion

Another, perhaps less appreciated point, is Mill’s insistence that law and public opinion may be equally burdensome. Law is not the only restriction, public opinion may be equally restrictive, and he even states that in England the yoke of opinion is heavier and the yoke of law lighter than in other European countries (Mill 1968: 128). Likewise, conduct is enforced not only by law, but also by opinion. Thus, Mill is dividing social pressure into the more formal law and the informal public opinion, understanding that the formal regulation is not the only one, and maybe not even the most important. Besides, he notes that human beings have a natural propensity for in-group intolerance, a large group or the majority in society most often requires the other to conform to the group’s truths. Thus, Mill is pointing to the dysfunctional pressure from public opinion and from the informal norms prevalent in every society. However, he does not seem to fully solve the problem of the yoke of public opinion: Whereas the yoke of the law can be dealt with, e. g. via constitutional protection, it is more difficult to see how the yoke of opinion can be dealt with. Mill sets up a normative ‘morality of public discussion’ including ‘giving merited honor to everyone, whatever opinion he may hold … exaggerating nothing to their discredit, keeping nothing back which tells … in their favor’ (Mill 1968: 171), but apparently he has no idea how to change conduct which is violating the morality of public discussion, let alone enforcing these important and in a sense indispensable norms.

4 Problems

There are various issues when going from the initial defense of free discussion to the contention that this is best protected by no legal restrictions. This paper will focus on the concept of language, but I shall briefly mention the other issues as they all seem to be somehow connected.

Firstly, Mill conceives law as restraints, indicating that if one wants to further freedom of speech, there ought to be as little law as possible. Obviously, this is not the whole story. Law may also function as a facilitator, and laws providing public service media, establishing a well-functioning media market and even public education are as important as – or even more important than – the laws pertaining to restricting speech. Even though there is no automatic connection between Mill’s conception of law and his perception of language, both suffer from a certain naivety, and there is a need to include more advanced conceptions when dealing with such precarious issues as freedom of speech. Seen in this light, H.L.A Hart’s 1961 criticism of the nineteenth century simplistic ‘gunman’ model of law, as if law is nothing but orders backed by threat, closely resembles John L. Austin’s criticism of the concept of language, as if language is nothing but propositions. Even so, this does not exclude that law, at times, may function as a restraint, e. g. when criminalizing defamatory speech. It is therefore still relevant to consider the construction of these constraints when dealing with freedom of speech.

Secondly, in a passing remark Mill excludes freedom of speech in ‘those backward states of society in which the race itself may be considered as in its nonage.’ For such societies, despotism is legitimate and probably even a prerequisite for the freedom to come. Liberty simply does not apply ‘anterior to the time when mankind has become capable of being improved by free and equal discussion’ (Mill 1968: 129, 130). One might dismiss this idea as an obsolete leftover from the imperialist past. Yet, the idea of backward and mature societies fits well with Mill’s insistence that democracy, including free discussion, has its prerequisites, such as an educated public, an ordered society and willingness to obedience (Mill 1861 in Malnes and Midgaard 2003: 165–167). Luckily, it is no longer standard to distinguish between the civilized and the uncivilized, but it remains a problem whether this distinction, so obvious to Mill, can simply be left out of the equation. The idea of a unilinear history has largely been abandoned; for various reasons it does not make much sense to conceive the history of mankind as passing through exactly the same stages of development, one of the reasons being that different civilizations have influenced and interfered differently with each other at different times. However, this does not entail that there is no development, and this does not imply that all societies at all times are ripe for exactly the same kind of regulation of freedom of speech. There is a difference between stable democracies and less stable, ^[2] and even more so between stable democracies with a well-educated population with experience in debating, working together despite differences in opinion etc., and societies with at times a heavy oppression.

Another outcome of abandoning the unilinear history is the fact that different societies may develop different solutions and regulations despite similarities as regards education, wealth etc. Mill himself indicates this by pointing out that in England the yoke of the public opinion is heavier than the yoke of the law in relation to freedom of speech, suggesting that the case is different in other European countries, especially in the Catholic countries, Mill seems to think. This raises the question of whether there is one best political philosophical solution to the issue of freedom of speech – which seems to be the assumption among freedom of speech philosophers as well as comparative lawyers of the functionalist school. Such an assumption is dubious and ought at least to be considered. In brief, I tend to think that there are better and worse solutions, but not one super-optimal solution fitting all societies, but rather differing ones, differing e. g. according to historical experiences, the legal system, trust in the state etc. ^[3]At any rate, when a difference is accepted – and Mill definitely recognized differences, only in terms of civilized versus non-civilized – it becomes an issue how to deal with such differences across borders. The easy, albeit unacceptable answer, is requiring others to abide by a universalized version of local or national norms. ^[4] The difficulty is, of course, to have internationally acceptable and well-functioning norms across different traditions and across different developments.

Does this, the idea of a unilinear history and nations divided into civilized and non-civilized, have any relation to the concept of language? I would argue in the affirmative. There is a striking similarity between the naive conceptions of law as restraints, development as unilinear and language as making propositions. In addition, having different legal systems with e. g. differing ways of dealing with freedom of speech resembles the fact that there are different language systems: Both are normative and both have developed local ways of dealing with all sorts of issues, without, I maintain, the option of reducing all solutions to one best solution.

Thirdly, Mill deals with discussions, ^[5] obviously having the sober discussion as the ideal type in which facts are brought forward no matter whether they support the defended opinion or the opposite and in which opposite views are not distorted. He is aware that this is far from always being the case, but he treats these aberrations (from the ideal that is) only as if they were individual shortcomings and not more systematic ways of dealing with public opinion. Mill hardly knew of professional formers of opinion, he hardly knew of paid political advertising with little if any content of discussion and with the wish of directly impressing certain views on the public opinion. Mill knew even less of systematized derision of groups, campaigns depicting certain ethnic or religious group as subhuman, as a disease or an evil that eventually must be extirpated. Indeed, Mill was well aware of the irrationality of human beings and of the public general opinion, but he did not treat this irrationality in any systematic way, apart from the unworkable and unacceptable idea of dividing the rational into two distinct geographical locations. He is still a child of the Enlightenment, seeing reason as a possible conqueror, being able to do away with irrationality and darkness. He did not see that freedom of speech could also include impressing the irrational, beyond discussion so to speak. Within the sphere of discussion, he could only defend freedom of speech, believing – and hoping – that reason and rationality would eventually win the day.

Fourthly, for Mill language is basically understood as producing ‘propositions’ in which something is held to be true or false (Mill 1882).Thus I may assert e. g. that gold is yellow (one of Mill’s examples) or that person X is an idiot, commodity Y is the best and cheapest and ethnic group Z are murderers and rapists (not Mill’s examples), propositions all of which may be true or false. Propositions basically make a claim by making a reference, ‘gold’ refers to a particular and distinguishable type of metal, ‘yellow’ to a particular and distinguishable type of color, ‘idiot’ to a particular and distinguishable type of human being, etc. The proposition may be more or less successful, gold may be yellow and X may be an idiot to a larger or lesser degree, and there are methods for providing evidence for the assertions. Mill, of course, goes further into the analysis, but for the present discussion it suffices to state that Mill conceives language as making propositions which can be more or less true or false.

This conception of language fits nicely with Mill’s defense of freedom of speech: Human beings are not likely to get everything right, and therefore we must be able to freely forward propositions in order for the truth to abound. And since the most convincing statements may be voiced in an insulting manner, there ought to be little or no restrictions of speech. Restrictions may prevent the participants from getting at the truth.

However, Mill’s conception of language is insufficient. Language is more than reference, and speakers do more than make propositions when they speak.

5 Improved conception of language

For the present purpose I have chosen two philosophers who have contributed considerably to an improved understanding of language, and whose ideas seem to have implications for freedom of speech, Gottlob Frege and John Langshaw Austin, respectively.

6 Briefly discussing with Butler

Judith Butler (1997) also engages John Austin when dealing with excitable speech and what she calls a politics of the performative. Without going into details of the entire work I want to dwell upon three aspects which I find interesting in relation to the consideration of the relationship between concepts of language and freedom of speech.

Firstly, Butler deals with the issue of linguistic vulnerability. As an answer to wounding words she does not recommend legal action since she finds courts and judgments to be equally violent, but rather advises us ^[8] to talk back. This, of course, makes sense. It is definitely worth considering whether problems, such as wounding words, should be dealt with by legislation and legal action, and probably even more so when we are dealing with speech. After all, freedom of speech is important and should not be abridged in order to counteract any undesirable side effect. Besides, there is a question of action and agency: Should the ones wounded by words take action themselves and talk back or leave the action to a third party. And in addition, at least sometimes there may be an option of dismantling the wounding effect. As an example, the male homosexuals in Denmark started to use a previously rather derogatory word, thus deflating wounding effect. With Frege, they made a new Sinn of being homosexuals.

On the critical side, however, I do not see why wounding is all that important. What matters about language is that it frames our minds and way of thinking. The verbal wounding is rather uninteresting, and the important thing about, say, defamation, is not that the defamed feels wounded but that his or her reputation may be damaged to an extent that the career is impeded and the social life ruined. I may be offended by being called a pedophile, but the real problem occurs when everybody conceives me as one. And if I as a professor am accused of taking bribes for giving good grades, the problem is not me getting grumpy but the threat to my career, my income and my social life. ^[9] The same goes for other grossly derogatory words. If they are used, say, any time a female politician publicly offers an opinion, the real problem is that this may turn out to be a de facto obstacle for democratic participation. Of course, it is uncomfortable to receive those kinds of words in your inbox but that is a smaller part of a much larger problem. I tend to think that Butler’s focus may relate to the implicit US law of torts, indicating damage and reparation, whereas the European approach tends to focus on protection of reputation as a part of private life. ^[10] On the other hand, the idea that defamation is about the right not to be grumpy is known in Denmark as well. A leading law professor has argued for the revocation of defamation laws for the exact reason that the law should not care about preventing people from feeling grumpy (Greve 1975, 2008). Such a perception is grossly underestimating the performative side of language.

Secondly, Butler is dealing with ‘hate speech’ without considering whether the term is appropriate and for what purposes. To me, ‘hate speech’ is a dubious concept and often a misnomer. Hate is neither a necessary or sufficient component of speech which may reasonably be considered to be legally or even criminally sanctioned. To hate someone is not criminal – not in my jurisdiction, at least – neither is it criminal to speak, ^[11]on the contrary. What matters are the larger effects of the words, the performative side to use an Austinian term, and as for the law, the challenge is to find a reasonable balance between protecting reputation etc. without infringing disproportionately on freedom of speech.

Thirdly, Butler seems to operate with US American law as the default position. I may, of course, be overdoing my point. Obviously, as an American she is writing about American cases such as R.A.V. v St. Paul and about the American Constitution’s first amendment. Yet, she is not writing a tract on American law or American legal philosophy, but a piece on excitable speech, apparently in general. Thus, while Mill is dividing the world into the civilized among whom freedom of speech should abound and the uncivilized not yet ripe for the blessings of freedom of speech (and a range of other rights), Butler seems to conceive everybody as civilized Americans, ^[12] facing problems of hate speech and interpretations of the US Constitution.

7 Bearings on freedom of speech

It is my contention that Fregeian and Austinian insights have bearings on our conception of freedom of speech. Obviously, particular legal or other normative solutions do not follow from the philosophical points. Even when we now know that words and sentences not only have reference but also Sinn, and even when we now know that language is not only – or even mainly – about making propositions, but also about doing, this acquired knowledge does not automatically entail specific ways of regulating speech.

Yet, Frege makes us aware that there is a difference between saying ‘Negro’, ‘black’ or ‘African American’, and that it is not a valid argument that since they refer to the same phenomenon, we can use the terms interchangeably. We cannot, because they have a different Sinn, a different sense, a different meaning. This does not mean, of course, that the use of any derogatory word ought to be criminalized, but it does mean that we need to be aware of the difference between reference and meaning. In my jurisdiction it is perfectly legal to discuss, say, an ethnic group‘s part of the prison population, whereas it may be criminally sanctioned to refer to the same group (same reference) as rapists, murderers, rats, vermin etc. Here a different meaning, a different Sinn is involved.

As for Austin, it appears that we may induce that absolute freedom of speech is not an option. Exactly because it is not possible to make a sharp distinction between propositions and performatives, between stating something and doing something, it is not possible to grant absolute freedom to speech but not to acts (the last of which nobody would dream of doing). The 1st Amendment of the US Constitution, ‘Congress shall make no law … abridging the freedom of speech…’ may look like an absolute freedom of speech, which of course it is not. The courts must produce a suitable doctrine in order to distinguish between constitutionally protected behavior and unprotected behavior, such as e. g. ‘the bad tendency’ and the ‘clear and present danger’ tests, the latter being more speech protecting than the former.

We can also learn that it is very difficult if not impossible to draw sharp distinctions when it comes to language. As an example, the European Court of Human Rights (ECtHR) initially made a matter-of-fact distinction between ‘statement of facts’ (propositions in Mill’ vocabulary) and ‘value judgments’, the last of which the Court conceived as subjective evaluations, i. e. words without reference. However, in Pedersen and Baadsgaard v Denmark, nine judges held certain words to be statements of facts whereas eight judges found the same words to be mere value judgments, demonstrating that the distinction does not always come easy. Had they read Austin, they would have been aware of the pitfalls. In later judgments the Court seems to be ready to admit that it is a continuum rather than a sharp distinction. ^[13]

Connected to this point, Austin also makes it clear that words and sentences have to be understood in a larger context. Consequently, it is not an option to classify certain words as prohibited if not true, because firstly words are more than propositions (which can be true or false), and secondly because the meaning depends on the situation. Obviously, there is a difference between greeting your pal with a “what’s up, old crook”, and charging an official for being a crook. The ECtHR is (now) quite aware of this fact, which at times may be very difficult to operationalize. ^[14]

Perhaps most importantly, Austin indicates that words and sentences have effects (the perlocutionary aspect) and that compared with physical acts there is a break in the chain when it comes to speech acts. This may hint that on the one hand we need to take effects into consideration – depicting someone as a KGB agent may have severe personal effects and depicting groups as vermin and rats may have severe consequences for persons belonging to that particular group. On the other hand, when adjudicating not all consequences should be ascribed to the speech act, neither the thinkable nor the real, and neither the intended nor the unintended. Caution is needed as to the remoteness of damages, and where exactly to draw the line is a mixed political, legislative and adjudicative activity. Ideally, knowledge of language ought to be included in the process, or better: a naïve and imperfect perception of language, such as Mill’s, ought not constitute the background for such important decisions. In addition, philosophy does not offer a solution, let alone the one and only solution. Local traditions, local language and knowledge of the language, local experience and local law must also be taken into consideration – in brief, civilization plays a part, albeit not in the rough Millian perception.

Thus, to conclude, I contend that it is possible to keep up a post-Millian defense for freedom of speech: Divergent and contrasting information and opinions are needed. It is highly unlikely that all possible information and opinions have been forwarded, and it is highly unlikely that the whole truth has been brought to light. And even if it were, counterarguments are still needed lest we fall into dogma and mental decay. However, as for the restraints, Mill’s naivety must be overcome: Speaking is not only making propositions, speaking is also always doing something and Sinn is also engaged.