This entry explores the topic of free speech. It starts with a general
discussion of freedom in relation to speech and then moves on to
examine one of the first and best defenses of free speech, based on
the harm principle. This provides a useful starting point for further
digressions on the subject. The discussion moves on from the harm
principle to assess the argument that speech can be limited because it
causes offense rather than direct harm. I then examine arguments that
suggest speech can be limited for reasons of democratic equality. I
finish with an examination of paternalistic and moralistic reasons
against protecting speech, and a reassessment of the harm principle.
The topic of free speech is one of the most contentious issues in
liberal societies. If liberty of expression is not highly valued, as
has often been the case, there is no problem; freedom of expression is
simply curtailed in favor of other values. It becomes a volatile issue
when it is highly valued because only then do the limitations placed
upon it become controversial. The first thing to note in any sensible
discussion of freedom of speech is that it will have to be limited.
Every society places some limits on the exercise of speech because it
always takes place within a context of competing values. In this
sense, Stanley Fish is correct when he says that there is no such
thing as free speech (in the sense of unlimited speech). Free speech
is simply a useful term to focus our attention on a particular form of
human interaction and the phrase is not meant to suggest that speech
should never be limited. One does not have to fully agree with Fish
when he says , “free speech in short, is not an independent
value but a political prize” (1994,102) but it is the case that
no society has existed where speech has not been limited to some
extent. Haworth (1998) makes a similar point when he suggests that a
right to freedom of speech is not something we have, not something we
own, in the same way as we possess arms and legs.
Alexander and Horton (1984) agree. They note that “speech”
encapsulates many different activities: speaking, writing, singing,
acting, burning flags, yelling on the street corner, advertising,
threats, slander and so on. One reason for thinking that speech is not
special simpiciter is that some of these forms of
communication are more important than others and hence require
different levels of protection. For example, the freedom to criticize
a government is generally thought to be more important than the
freedom of an artist to offend her audience. If two speech acts clash
(when yelling prevents a political speech) a decision has to be made
to prioritize one over the other, which means that there can be no
unlimited right to free speech. For example, Alexander and Horton
(1984) claim that arguments defending speech on democratic grounds
have many parts. One is a claim that the public needs a great deal of
information in order to make informed decisions. Another is that
because government is the servant of the people, it should not be
allowed to censor them. Such arguments show that one of the main
reasons for justifying free speech (political speech) is important,
not for its own sake but because it allows us to exercise another
important value (democracy). Whatever reasons we offer to protect
speech can also be used to show why some speech is not special. If
speech is defended because it promotes autonomy, we no longer have
grounds for protecting speech acts that undermine this value. If our
defence of speech is that it is crucial to a well-functioning
democracy, we have no reason to defend speech that is irrelevant to,
or undermines, this goal. And if we agree with John Stuart Mill (1978)
that speech should be protected because it leads to the truth, there
seems no reason to protect the speech of anti-vaccers or creationists.
Speech is important because we are socially situated and it makes
little sense to say that Robinson Crusoe has a right to free speech.
It only becomes necessary to talk of such a right within a social
setting, and appeals to an abstract and absolute right to free speech
hinder rather than help the debate. At a minimum, speech will have to
be limited for the sake of order. If we all speak at once, we end up
with an incoherent noise. Without some rules and procedures we cannot
have a conversation at all and consequently speech has to be limited
by protocols of basic civility.
It is true that many human rights documents give a prominent place to
the right to speech and conscience, but such documents also place
limits on what can be said because of the harm and offense that
unlimited speech can cause, (I will discuss this in more detail
later). Outside of the United States of America speech does not tend
to have a specially protected status and it has to compete with other
rights claims for our allegiance. John Stuart Mill, one of the great
defenders of free speech, summarized these points in On
Liberty, where he suggests that a struggle always takes place
between the competing demands of authority and liberty. He claimed
that we cannot have the latter without the former:
All that makes existence valuable to anyone depends on the
enforcement of restraints upon the actions of other people. Some rules
of conduct, therefore, must be imposed—by law in the first
place, and by opinion on many things which are not fit subjects for
the operation of law. (1978, 5)
The task, therefore, is not to argue for an unlimited domain of free
speech; such a concept cannot be defended. Instead, we need to decide
how much value we place on speech in relation to other important
ideals such as privacy, security, democratic equality and the
prevention of harm and there is nothing inherent to speech that
suggests it must always win out in competition with these values.
Speech is part of a package deal of social goods: “speech, in
short, is never a value in and of itself but is always produced within
the precincts of some assumed conception of the good” (Fish,
1994, 104). In this essay, I will examine some conceptions of the good
that are deemed to be acceptable limitations on speech. I will start
with the harm principle and then move on to other more encompassing
arguments for limiting speech.
Before we do this, however, the reader might wish to disagree with the
above claims and warn of the dangers of the “slippery
slope.” As Frederick Schauer (1985) has demonstrated, slippery
slope arguments make the claim that a current acceptable change (he
calls this the instant case) to the status quo regarding speech will
lead to some intolerable future state of affairs (what he calls the
danger case) once the instant case prohibiting speech is introduced.
The assumption is that the instant case is acceptable; otherwise it
would be critiqued in its own right. The complaint is that a change
from the status quo to the instant case will lead to unwanted future
limitations on speech and should be avoided (even if a change to the
instant case would be immediately desirable). The slippery slope
argument has to make a clear distinction between the instant and the
danger case. If the former was part of the latter then it is not a
slippery slope argument but simply an assertion about the unwarranted
breadth of the instant case. The claim being made is that a change to
an acceptable instant case that is distinct from the danger case
should nevertheless be prohibited because a change from the status quo
to the instant case will necessarily transport us to the danger case.
As Schuer says this is not very compelling because it needs to be
demonstrated, rather than merely stated, that the move from the status
quo is so much more likely to lead to the danger case. Part of the
problem is that slippery slope arguments are often presented in a way
that suggests we can be on or off the slope. In fact, no such choice
exists: we are necessarily on the slope whether we like it or
not, and the task is always to decide how far up or down we choose to
go, not whether we should step off the slope altogether. We need to
keep in mind that the slippery-slope claim is not that the proposed
instant case will lead to minor changes in the future, but that a
small change now will have drastic and tyrannical consequences. The
slippery-slope argument seems to suggest that the instant case is so
flawed that any change to it from the status quo (which again, is a
position already on the slope) puts us in imminent threat of sliding
into the danger case. Unfortunately, the causal mechanisms for how
this must necessarily happen are usually unspecified. Anyone
making such claims should be willing to demonstrate how this unlikely
event will happen before being taken seriously. Such a person is not
simply advocating caution; she is claiming that there is an imminent
risk of moving from an acceptable instant case to an unacceptable
danger case. This is not to say that slippage cannot occur. One
safeguard against this is to be as precise as possible in our use of
language. If harm to others is our preferred stopping point on the
slope, we need to specify in clear terms what counts as harm and what
does not. Sometimes we will fail in this task, but precision puts
brakes on the instant case and limits its capacity for sliding down
the slope.
Those who support the slippery slope argument tend to make the claim
that the inevitable consequence of limiting speech is a slide into
censorship and tyranny. It is worth noting, however, that the slippery
slope argument can be used to make the opposite point; one could argue
that we should not allow any removal of government interventions (on
speech or any other type of freedom) because once we do we are on the
slippery slope to anarchy, the state of nature, and a life that Hobbes
described in Leviathan as “solitary, poore, nasty,
brutish, and short” (1968, 186).
It is possible that some
limits on speech might, over time, lead to further
restrictions—but they might not. And if they do, those
limitations might also be justified. The main point is that once we
abandon the incoherent position that there should be no limits on
speech, we have to make controversial decisions about what can and
cannot be expressed; this comes along with the territory of living
together in communities.
Another thing to note before we engage with specific arguments for
limiting speech is that we are in fact free to speak as we like.
Hence, freedom of speech differs from some other types of free action.
If the government wants to prevent citizens engaging in certain
actions, riding motor bikes for example, it can limit their freedom to
do so by making sure that such vehicles are no longer available;
current bikes could be destroyed and a ban can be placed on future
imports. Freedom of speech is a different case. A government can limit
some forms of free expression by banning books, plays, films etc. but
it cannot make it impossible to say certain things. The only thing it
can do is punish people after they have spoken. This means
that we are free to speak in a way that we are not free to ride
outlawed motorbikes. This is an important point; if we insist that
legal prohibitions remove freedom then we have to hold the
incoherent position that a person was unfree at the very moment she
performed a speech act. The government would have to remove our vocal
cords for us to be unfree in the same way as the motorcyclist is
A more persuasive analysis suggests that the threat of a sanction
makes it more difficult and potentially more costly to exercise our
freedom of speech. Such sanctions take two major forms. The first, and
most serious, is legal punishment by the state, which usually consists
of a financial penalty, but can stretch to imprisonment (which then,
of course, further restricts the persons free speech). The second
threat of sanction comes from social disapprobation. People will often
refrain from making public statements because they fear the ridicule
and moral outrage of others. For example, one could expect to be
publicly condemned if one made racist comments during a public lecture
at a university. Usually it is the first type of sanction that catches
our attention but, as we will see, John Stuart Mill provides a strong
warning about the chilling effect of the latter form of social
We seem to have reached a paradoxical position. I started by claiming
that there can be no such thing as a pure form of free speech: now I
seem to be arguing that we are, in fact, free to say anything we like.
The paradox is resolved by thinking of free speech in the following
terms. I am, indeed, free to say (but not necessarily to publish) what
I like, but the state and other individuals can sometimes make that
freedom more or less costly to exercise. This leads to the conclusion
that we can attempt to regulate speech, but we cannot prevent it if a
person is undeterred by the threat of sanction. The issue, therefore,
boils down to assessing how cumbersome we wish to make it for people
to say certain things. I have already suggested that all societies do
(correctly) make some speech more costly than others. If the reader
doubts this, it might be worth considering what life would be like
with no sanctions on libelous statements, child pornography,
advertising content, and releasing state secrets. The list could go
The conclusion to be drawn is that the problem we face is deciding
where, not whether, to place limits on speech, and the next sections
look at some possible solutions to this puzzle.
Given that Mill presented one of the first, and still perhaps the most
famous liberal defense of free speech, I will focus on his arguments
in this essay and use them as a springboard for a more general
discussion of free expression. In the footnote at the beginning of
Chapter II of On Liberty, Mill makes a very bold statement:
If the arguments of the present chapter are of any validity, there
ought to exist the fullest liberty of professing and discussing, as a
matter of ethical conviction, any doctrine, however immoral it may be
considered. (1978, 15)
This is a very strong defense of free speech; Mill tells us that
any doctrine should be allowed the light of day no matter
how immoral it may seem to everyone else. And Mill does mean
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. (1978, 16)
Such liberty should exist with every subject matter so that we have
“absolute freedom of opinion and sentiment on all subjects,
practical or speculative, scientific, moral or theological”
(1978, 11). Mill claims that the fullest liberty of expression is
required to push our arguments to their logical limits, rather than
the limits of social embarrassment. Such liberty of expression is
necessary, he suggests, for the dignity of persons. If liberty of
expression is stifled, the price paid is “a sort of intellectual
pacification” that sacrifices “ the entire moral courage of the human
mind” (1978, 31).
These are powerful claims for freedom of speech, but as I noted above,
Mill also suggests that we need some rules of conduct to regulate the
actions of members of a political community. The limitation he places
on free expression is “one very simple principle” (1978,
9), now usually referred to as the harm principle, which states
…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. (1978, 9)
There is a great deal of debate about what Mill had in mind when he
referred to harm; for the purposes of this essay he will be taken to
mean that an action has to directly and in the first instance invade
the rights of a person (Mill himself uses the term rights, despite
basing the arguments in the book on the principle of utility). The
limits on free speech will be very narrow because it is difficult to
support the claim that most speech causes harm to the rights of
others. This is the position staked out by Mill in the first two
chapters of On Liberty and it is a good starting point for a
discussion of free speech because it is hard to imagine a more liberal
position. Liberals are usually willing to contemplate limiting speech
once it can be demonstrated that it does invade the rights of
If we accept Mill’s argument we need to ask “what types of
speech, if any, cause harm?” Once we can answer this question,
we have found the appropriate limits to free expression. The example
Mill uses is in reference to corn dealers: he suggests that it is
acceptable to claim that corn dealers starve the poor if such a view
is expressed in print. It is not acceptable to make such statements to
an angry mob, ready to explode, that has gathered outside the house of
the corn dealer. The difference between the two is that the latter is
an expression “such as to constitute…a positive
instigation to some mischievous act,” (1978, 53), namely, to
place the rights, and possibly the life, of the corn dealer in danger.
As Daniel Jacobson (2000) notes, it is important to remember that Mill
will not sanction limits to free speech simply because someone is
harmed. For example, the corn dealer may suffer severe financial
hardship if he is accused of starving the poor. Mill distinguishes
between legitimate and illegitimate harm, and it is only when speech
causes a direct and clear violation of rights that it can be limited.
The fact that Mill does not count accusations of starving the poor as
causing illegitimate harm to the rights of corn dealers suggests he
wished to apply the harm principle sparingly. Other examples where the
harm principle may apply include libel laws, blackmail, advertising
blatant untruths about commercial products, advertising dangerous
products to children (e.g. cigarettes), and securing truth in
contracts. In most of these cases, it is possible to show that harm
can be caused and that rights can be violated.
There are other instances when the harm principle has been invoked but
where it is more difficult to demonstrate that rights have been
violated. Perhaps the most obvious example is the debate over
pornography. As Feinberg notes in Offense to Others: the Moral
Limits of the Criminal Law, most attacks on pornography up to the
1970s were from social conservatives who found such material to be
immoral and obscene. This type of argument has died away in recent
times and the case against pornography has been taken up by some
feminists who often distinguish between erotica, which is acceptable,
and pornography, which is not, because it is claimed it degrades,
harms, and endangers the lives of women. The harm principle can be
invoked against pornography if it can be demonstrated that it violates
the rights of women.
This is an approach taken by Catherine MacKinnon (1987). She takes
seriously the distinction between pornography and erotica. Erotica
might be explicit and create sexual arousal, but neither is grounds
for complaint. Pornography would not come under attack if it did the
same thing as erotica; the complaint is that it portrays women in a
manner that harms them.
When pornography involves young children, most people accept that it
should be prohibited because it harms persons under the age of
consent (although the principle would not necessarily rule out people
over the age of consent from portraying minors). It has proved more
difficult to make the same claim for consenting adults. It is
difficult to know if the people who appear in books, magazines, films,
videos and on the internet are being physically harmed. If they are we
then need to show why this is sufficiently different from other forms
of harmful employment that is not prohibited, such as hard manual
labour, or very dangerous jobs. Much of the work in pornography seems
to be demeaning and unpleasant but the same can be said for many forms
of work and again it is unclear why the harm principle can be used to
single out pornography. MacKinnon’s (1987) claim that women who make a
living through pornography are sexual slaves seems to exaggerate the
case. If conditions in the pornography industry are particularly bad,
stronger regulation rather than prohibition might be a better option,
particularly as the latter will not make the industry go away.
It is also difficult to demonstrate that pornography results in harm
to women as a whole. Very few people would deny that violence against
women is abhorrent and an all too common feature of our society, but
how much of this is caused by pornography? MacKinnon, Andrea Dworkin,
(1981) and many others, have attempted to show a causal link but this
has proven challenging because one needs to show that a person who
would not rape, batter or otherwise violate the rights of women was
caused to do so through exposure to pornography. Caroline West
provides a useful overview of the literature and suggests that even
though pornography might not dispose most men to rape, it might make
it more likely for those men who are already so inclined. She uses the
analogy of smoking. We have good grounds for saying that smoking makes
cancer more likely even though smoking is neither a necessary nor
sufficient condition for causing cancer. One possible problem with
this analogy is that we have very powerful evidence that smoking does
significantly increase the possibility of cancer; the evidence
suggesting that viewing pornography leads men (already inclined) to
rape women is not as robust.
If pornographers were exhorting their readers to commit violence and
rape, the case for prohibition would be much stronger, but they tend
not to do this, just as films that depict murder do not actively
incite the audience to mimic what they see on the screen. For the sake
of argument let us grant that the consumption of pornography does lead
some men to commit acts of violence. Such a concession might not prove
to be decisive. The harm principle might be a necessary, but it is not
a sufficient reason for censorship. If pornography causes a small
percentage of men to act violently we still need an argument for why
the liberty of all consumers of pornography (men and women) has to be
curtailed because of the violent actions of a few. We have
overwhelming evidence that consuming alcohol causes a lot of violence
(against women and men) but this does not mean that alcohol should be
prohibited. Very few people reach this conclusion despite the clarity
of the evidence. Further questions need to be answered before a ban is
justified. How many people are harmed? What is the frequency of the
harm? How strong is the evidence that A is causing B? Would
prohibition limit the harm and if so, by how much? Would censorship
cause problems greater than the harm it is meant to negate? Can the
harmful effects be prevented by measures other than prohibition?
There are other non-physical harms that also have to be taken into
consideration. MacKinnon argues that pornography causes harm because
it exploits, oppresses, subordinates and undermines the civil rights
of women, including their right to free speech. A permissive policy on
pornography has the effect of prioritizing the right to speech of
pornographers over the right to speech of women. MacKinnon’s claim is
that pornography silences women because it presents them as inferior
beings and sex objects who are not to be taken seriously. Even if
pornography does not cause violence, it still leads to discrimination,
domination and rights violations. She also suggests that because
pornography offers a misleading and derogatory view of women, it is
libelous. Along with Andrea Dworkin, MacKinnon drafted a Minneapolis
Council Ordinance in 1983 that allowed women to take civil action
against pornographers. They defined pornography as:
…the graphic sexually explicit subordination of women through
pictures or words that also includes women dehumanized as sexual
objects, things, or commodities; enjoying pain or humiliation or rape;
being tied up, cut up, mutilated, bruised, or physically hurt; in
postures of sexual submission or servility or display; reduced to body
parts, penetrated by objects or animals, or presented in scenarios of
degradation, injury, torture; shown as filthy or inferior; bleeding,
bruised or hurt in a context which makes these conditions
sexual (1987, 176).
Such arguments have so far not led to the prohibition of pornography
(which was not the intent of the Ordinance) and many liberals remain
unconvinced. One reason that some doubt MacKinnon’s claims is that the
last twenty years have seen an explosion of pornography on the
internet without a concurrent erosion of women’s rights. If those
arguing that pornography causes harm are right, we should expect to
see a large increase in physical abuse against women and a hefty
decrease in their civil rights, employment in the professions, and
positions in higher education. The evidence does not seem to show this
and social conditions for women today are better than 30 years ago
when pornography was less prevalent. What does seem to be reasonably
clear, at least in the USA, is that the increased consumption of
pornography over the last 20 years has coincided with a reduction in
violent crime against women, including rape. If we return to West’s
smoking analogy, we would have to rethink our view that smoking causes
cancer if a large increase in smokers did not translate into a
comparable increase in lung cancer.
The matter remains unsettled, and the lives of women might be
significantly better if pornography was not around, but so far it has
proven difficult to justify limiting pornography by way of the harm
principle. It is important to remember that we are currently examining
this issue from the perspective of Mill’s formulation of the harm
principle and only speech that directly violates rights
should be banned. Finding pornography offensive, obscene or outrageous
is not sufficient grounds for censorship. Nor does Mill’s principle
allow prohibition because pornography harms the viewer. The harm
principle is there to prevent other-regarding not self-regarding harm.
Overall, no one has mounted a compelling case (at least as far as
legislators and judges are concerned) for banning pornography (except
in the case of minors) based on the concept of harm formulated by
Another difficult case is hate speech. Most liberal democracies have
limitations on hate speech, but it is debatable whether these can be
justified by the harm principle as formulated by Mill. One would have
to show that such speech violated rights, directly and in the first
instance. I am interested here in hate speech that does not advocate
violence against a group or individual because such speech would be
captured by Mill’s harm principle. The Public Order Act 1986 in the
U.K. does not require such a stringent barrier as the harm principle
to prohibit speech. The Act states that “A person is guilty of
an offence if he …displays any writing, sign or other visible
representation which is threatening, abusive or insulting, within the
hearing or sight of a person likely to be caused harassment, alarm or
There have been several prosecutions in the U.K. that would not have
happened if the harm principle governed “absolutely the dealings
of society with the individual”(Mill,1978, 68). In 2001
evangelist Harry Hammond was prosecuted for the following statements:
“Jesus Gives Peace, Jesus is Alive, Stop Immorality, Stop
Homosexuality, Stop Lesbianism, Jesus is Lord”. For his sins he
was fined 300 pounds and made to pay 395 pounds in costs. In 2010,
Harry Taylor left anti-religious cartoons in the prayer-room of
Liverpool’s John Lennon Airport. The airport chaplain was
“insulted, offended, and alarmed” by the cartoons and
called the police. Taylor was prosecuted and received a six-month
suspended sentence. Barry Thew wore a t-shirt hours after two women
police officers were murdered near Manchester in 2012. The front of
the shirt had the slogan “One less pig, perfect justice,”
and on the back was written “Kill a cop for fun”. He
admitted a Section 4A Public Order Offence and was sentenced to 4
months jail. Also in 2012, Liam Stacey took to twitter to mock a black
professional football player who collapsed during a match. He then
proceeded to racially abuse people who responded negatively to his
tweet. He was sentenced to 56 days in jail. This case provoked
significant commentary, most of it taking the form of slippery-slope
claims that the decision would inevitably lead to Britain becoming a
totalitarian state. The most recent (June 2016) case to receive public
attention involves Paul Gascoigne, the former English football star,
who has been charged with racially aggravated abuse after commenting,
whilst on stage, that he could only make out a black man standing in a
dark corner of the room when he smiled. It is doubtful that any of
these examples would be captured by Mill’s harm principle.
In Australia, Section 18C of the Racial Discrimination Act 1975 states
that “It is unlawful for a person to do an act, otherwise than
in private, If: (a) the act is reasonably likely in all the
circumstances to offend, insult, humiliate or intimidate another
person or group of people, and (b) the act is done because of race,
colour or national or ethnic origin”. The most prominent person
prosecuted under the Act is Andrew Bolt, a conservative political
commentator, who was found guilty of racially vilifying nine
aboriginal persons in newspaper articles in 2011. He suggested that
the nine people had identified as aboriginal, despite having fair
skin, for their own professional advantage. The case prompted the Tony
Abbott led Liberal government into a failed attempt to change the
It should be noted that Section 18C is qualified by Section 18D (often
ignored in the backlash against the Bolt decision). 18D says that
…section 18C does not render unlawful anything said or done
reasonably and in good faith: (a) in the performance, exhibition or
distribution of an artistic work; or (b) in the course of any
statement, publication, discussion or debate made or held for any
genuine academic, artistic or scientific purpose or any other genuine
purpose in the public interest; or (c) in the making or publishing:
(i) a fair and accurate report of any event or matter of public
interest; or (ii) a fair comment on any matter of public interest if
the comment is an expression of a genuine belief held by the person
making the comment…
It is clear that these qualifications remove some of the teeth from
Section 18C. As long as the statements are made artistically and/or in
good faith, for example, they are immune from prosecution. The
conclusion of the judge in the Bolt case was that none of the Section
18D exemptions applied in his case. Even with these qualifications in
place, however, it seems that the Racial Discrimination Act would
still be ruled out by Mill’s harm principle which seems to allow
people to offend, insult, and humiliate (although perhaps not
intimidate) regardless of the motivation of the speaker.
The United States, precisely because it fits most closely with Mill’s
principle, is an outlier amongst liberal democracies when it comes to
hate speech. The most famous example of this is the Nazi march through
Skokie, Illinois, something that would not be allowed in many other
liberal democracies. The intention was not to engage in political
speech at all, but simply to march through a predominantly Jewish
community dressed in storm trooper uniforms and wearing swastikas
(although the Illinois Supreme Court interpreted the wearing of
swastikas as “symbolic political speech”). It is clear
that many people, especially those who lived in Skokie, were outraged
and offended by the march, but were they harmed? There was no plan to
cause physical injury and the marchers did not intend to damage
The main argument for prohibiting the Skokie march, based on
considerations of harm, was that the march would incite a riot, thus
putting the marchers in danger. The problem with this argument is that
the focal point is the potential harm to the speakers and not the harm
done to those who are the subject of the hate. To ban speech for this
reason, i.e., for the good of the speaker, tends to undermine the
basic right to free speech in the first place. If we turn our
attention to members of the local community, we might want to claim
that they were psychologically harmed by the march. This is much more
difficult to demonstrate than harm to a person’s legal rights. It
seems, therefore, that Mill’s argument does not allow for state
intervention in this case. If we base our defense of speech on Mill’s
principle we will have very few prohibitions. It is only when we can
show direct harm to rights, which will almost always mean when an
attack is made against a specific individual or a small group of
persons, that it is legitimate to impose a sanction.
One response is to suggest that the harm principle can be defined less
stringently. Jeremy Waldron (2012) has made a recent attempt to do
this. He draws our attention to the visual impact of hate speech
through posters and signs displayed in public. Waldron argues that the
harm in hate speech (the title of his book) is that it compromises the
dignity of those under attack. A society where such images proliferate
makes life exceedingly difficult for those targeted by hate speech.
Waldron suggests that the people engaged in hate speech are saying
“[t]he time for your degradation and your exclusion by the society
that presently shelters you is fast approaching” (2012, 96). He
claims that prohibiting such messages assures all people that they are
welcome members of the community.
Waldron does not want to use hate speech legislation to punish those
who hold hateful thoughts and attitudes. The goal is not to engage in
thought control but to prevent harm to the social standing of certain
groups in society. Liberal democratic societies are founded on ideas
of equality and dignity and these are damaged by hate speech. Given
this, Waldron wonders why we even need to debate the usefulness of
hate speech. Mill, for example, argued that we should allow speech of
this type so that our ideas do not fall into the “slumber of a
decided opinion” (1978, 41). Waldron doubts that we require hate
speech to prevent such an outcome.
As we have seen, Waldron is making a harm based argument but his
threshold for what counts as harm is lower than Mill’s. He needs to
convince us that an attack on a person’s dignity constitutes a
significant harm. My dignity might often be bruised by colleagues, for
example, but this does not necessarily show that I have been harmed.
Perhaps it is only when an attack on dignity is equivalent to threats
of physical abuse that it counts as a reason for limiting speech.
Waldron does not offer a lot of evidence that a permissive attitude to
hate speech, at least in liberal democracies, does cause significant
harm. There is no specific hate speech regulation in the United
States, for example, but it is not clear that more harm occurs there
than in other liberal democracies.
David Boonin (2011) is not convinced that there is a need for special
hate speech legislation. He claims that hate speech does not fit
within the regular categories of speech that can be prohibited. Even
if he can be persuaded that it does fit, he still thinks special hate
speech laws are not required because existing legislation will capture
the offending speech. I will examine one example he uses to make his
point. Boonin argues that threatening speech already sits within the
category of speech that is rightfully prohibited. He suggests,
however, that hate speech does not fall within this category because a
significant amount of hate speech is not directly threatening. A group
of black men, for example, will not be threatened by a racially
abusive elderly white woman. He argues that this example, and others
like it, show why a blanket ban on all hate speech on the grounds that
it is threatening cannot be justified.
Nor is it likely, he suggests, that racist attacks by frail old ladies
will contribute to an atmosphere of danger. This argument might be
less persuasive. Mill’s use of the corn dealer example demonstrates
how the use of language can incite violence regardless of who is
speaking. But Mill’s example also shows that a blanket ban would
still be unwarranted because it allows incendiary statements to be
made about corn dealers under controlled conditions.
Boonin’s argument does not rest here. If it really does turn out to
be the case that all hate speech is threatening in the appropriate
sense, this still does not justify special hate speech laws because
there is already legislation in place prohibiting threatening
language. Boonin is opposed to banning hate speech because it is
hateful not because it is threatening. He claims that the argument for
special hate speech laws is “impaled on the horns of a dilemma:
either the appeal is unconvincing because not all forms of hate speech
are threatening, or it is unnecessary precisely because all forms of
hate speech are threatening and are therefore already prohibited”
(2011, 213). Boonin uses the same strategy with regard to other
reasons, such as “fighting words”, for banning hate speech; they all
find themselves impaled on the horns of the same dilemma.
The arguments of Waldron and Boonin seem to be a long way apart and
the latter suggests that anyone who argues for hate speech laws is
taking an extreme position. There is, however, a lot of overlap
between the two, particularly as both focus on harm, and neither wants
to censor hate speech simply because it is offensive. This becomes
clearer if we take a suggestion offered by Waldron. At one point in
his book he ponders whether it might be advantageous to abandon the
term “hate speech” altogether. Such a move goes a long way to
reconciling the arguments of Waldron and Boonin. Both authors agree
that prohibition is acceptable when speech is threatening; they
disagree on what counts as a harmful threat. Waldron thinks most forms
of racial abuse qualify whereas Boonin is more circumspect. But the
disagreement between the two is about what causes harm rather than any
major philosophical difference about the appropriate limits on speech.
If both agree that a threat constitutes a significant harm, then both
will support censorship. This still leaves lots of room for
disagreement, particularly as we are now more aware than was Mill of
psychological as well as physical harm. I cannot delve into the topic
here except to say that if we expand the harm principle from the
physical to the mental realm, more options might become available for
prohibiting hate speech and pornography.
There are two basic responses to the harm principle. One is that it is
too narrow; the other is that it is too broad. This latter view is not
often expressed because, as already noted, most people think that free
speech should be limited if it causes illegitimate harm. George Kateb
(1996), however, has made an interesting argument that runs as
follows. If we want to limit speech because it causes harm, we will
have to ban a lot of political speech. Most of it is useless, a lot of
it is offensive, and some of it causes harm because it is deceitful
and aimed at discrediting specific groups. It also undermines
democratic citizenship and stirs up nationalism and jingoism, which
results in harm to citizens of other countries. Even worse than
political speech, according to Kateb, is religious speech. He claims
that a lot of religious speech is hateful, useless, dishonest, and
foments war, bigotry and fundamentalism. It also creates bad
self-image and feelings of guilt that can haunt persons throughout
their lives. Pornography and hate speech, he claims, cause nowhere
near as much harm as political and religious speech. As we rightly do
not want to ban political and religious speech, Kateb claims to have
demonstrated that the harm principle casts the net too far. His
solution is to abandon the principle in favor of almost unlimited
This is a powerful argument, but there seem to be at least two
problems. The first is that the harm principle would actually allow
religious and political speech for the same reasons that it allows
most pornography and hate speech, namely that it is not possible to
demonstrate that such speech does cause direct harm to rights. I doubt
that Mill would support using his arguments about harm to ban
political and religious speech. The second problem for Kateb is that
if he is right that such speech does cause harm by violating rights,
we now have powerful reasons for limiting political and religious
speech. If Kateb’s argument is sound he has shown that harm is more
extensive than we might have thought; he has not demonstrated that the
harm principle is invalid.
The other response to the harm principle is that it does not reach far
enough. One of the most impressive arguments for this position comes
from Joel Feinberg who suggests that the harm principle cannot
shoulder all of the work necessary for a principle of free speech. In
some instances, Feinberg suggests, we also need an offense
principle that can guide public censure. The basic idea is that
the harm principle sets the bar too high and that we can legitimately
prohibit some forms of expression because they are very offensive.
Offending is less serious than harming so any penalties imposed should
not be severe. As Feinberg notes, this has not always been the case
and he cites a number of instances in the U.S. where penalties for
“offensive” acts like sodomy and consensual incest have ranged from
twenty years imprisonment to the death penalty. Feinberg’s principle
reads as follows: “it is always a good reason in support of a
proposed criminal prohibition that it would probably be an effective
way of preventing serious offense…to persons other than the actor,
and that it is probably a necessary means to that end…The principle
asserts, in effect, that the prevention of offensive conduct is
properly the state’s business” (1985, 1).
Such a principle is hard to apply because many people take offense as
the result of an overly sensitive disposition, or worse, because of
bigotry and unjustified prejudice. A further difficulty is that some
people can be deeply offended by statements that others find mildly
amusing. The furore over the Danish cartoons brings this starkly to
the fore. Despite the difficulty of applying a standard of this kind,
something like the offense principle operates widely in liberal
democracies where citizens are penalized for a variety of activities,
including speech, that would escape prosecution under the harm
principle. Wandering around the local shopping mall naked, or engaging
in sexual acts in public places are two obvious examples. Given the
specific nature of this essay, I will not delve into the issue of
offensive behavior in all its manifestations, and I will limit the
discussion to offensive forms of speech. Feinberg suggests that many
factors need to be taken into account when deciding whether speech can
be limited by the offense principle. These include the extent,
duration and social value of the speech, the ease with which it can be
avoided, the motives of the speaker, the number of people offended,
the intensity of the offense, and the general interest of the
How does the offense principle help us deal with the issue of erotica?
Given the above criteria, Feinberg argues that books should never be
banned because the offensive material is easy to avoid. If one is
unaware of the content and should become offended in the course of
reading the text, the solution is simple-close the book. A similar
argument would be applied to erotic films. The French film Baise-Moi
was in essence banned in Australia in 2002 because of its supposed
offensive material (it was denied a rating which meant that it could
not be shown in cinemas). It would seem, however, that the offense
principle outlined by Feinberg would not permit such prohibition
because it is very easy to avoid being offended by the film. It should
also be legal to advertise the film, but some limits could be placed
on the content of the advertisement so that sexually explicit material
is not placed on billboards in public places (because these are not
easily avoidable). At first glance it might seem strange to have a
more stringent speech code for advertisements than for the thing being
advertised; the harm principle would not provide the grounds for such
a distinction, but it is a logical conclusion of the offense
What of pornography i.e. material that is offensive because of its
extremely violent or degrading content? In this case the offense is
more profound: simply knowing that such material exists is enough to
deeply offend many people. The difficulty here is that bare knowledge,
i.e., being offended by knowing that something exists or is taking
place, is not as serious as being offended by something that one does
not like and that one cannot escape. If we allow that films
should be banned because some people are offended, even when they do
not have to view them, consistency demands that we allow the
possibility of prohibiting many forms of expression. A lot of people
find strong attacks on religion, or t.v. shows by religious
fundamentalists deeply offensive. Feinberg argues that even though
some forms of pornography are profoundly offensive to many people,
they should not be prohibited on these grounds.
Hate speech causes profound offense. The discomfort caused to the
targets of such attacks cannot be shrugged off easily. As with violent
pornography, the offense that is caused by the march through Skokie
cannot be avoided simply by staying off the streets because offense is
taken over the bare knowledge that the march is taking place. As we
have seen, however, bare knowledge does not seem sufficient grounds
for prohibition. But in respect to some of the other factors regarding
offensive speech mentioned above, Feinberg suggests that the march
through Skokie does not do very well: the social value of the speech
seems to be marginal, the number of people offended will be large, and
it is difficult to see how it is in the interests of the community.
These reasons also hold for violent pornography which Feinberg
suggests should not be prohibited for reasons of offense.
A key difference, however, is the intensity of the offense; it is
particularly acute with hate speech because it is aimed at a
relatively small and specific audience. The motivations of the
speakers in the Skokie example seemed to be to incite fear and hatred
and to directly insult members of the community through the use of
Nazi symbols. Nor, according to Feinberg, was there any political
content to the speech. The distinction between violent pornography and
the Skokie example of hate speech is that a particular group of people
were targeted and the message of hate was paraded in such a way that
it could not be easily avoided. It is for these reasons that Feinberg
suggests hate speech can be limited by the offense principle.
He also claims that when fighting words are used to provoke people who
are prevented by law from using a fighting response, the offense is
profound enough to allow for prohibition. If pornographers engaged in
the same behaviour and paraded through neighborhoods where they were
likely to meet great resistance and cause profound offense, they too
should be prevented from doing so. It is clear, therefore, that the
crucial component of the offense principle is whether the offense can
be avoided. Feinberg’s principle means that many forms of hate speech
will still be allowed if the offense is easily avoidable. It still
allows Nazis to meet in private places, or even in public ones that
are easily bypassed. Advertisements for such meetings can be edited
(because they are less easy to avoid) but should not be banned. It
seems Feinberg thinks that hate speech does not, in and of itself,
cause direct harm to the rights of the targeted group (he is not
claiming that offence equals harm) and he would be troubled by some of
the prohibitions on speech in the U.K. and Australia.
Very few, if any, liberal democracies are willing to support the
Millian view that only speech causing direct harm to rights should be
prohibited. Most support some form of the offense principle. Some
liberal philosophers are willing to extend the realm of state
interference further and argue that hate speech should be banned even
if it does not cause harm or unavoidable offense. The reason it should
be banned is that it is inconsistent with the underlying values of
liberal democracy to brand some citizens as inferior on the grounds of
race, religion, gender or sexual orientation. The same applies to
pornography; it should be prevented because it is incompatible with
democratic citizenship to portray women as submissive sexual objects,
who seem to enjoy being violently mistreated. Rae Langton, for
example, starts from the liberal premise of equal concern and respect
and concludes that it is justifiable to remove certain speech
protections for pornographers. She avoids basing her argument on harm:
“If, for example, there were conclusive evidence linking
pornography to violence, one could simply justify a prohibitive
strategy on the basis of the harm principle. However, the prohibitive
arguments advanced in this article do not require empirical premises
as strong as this…they rely instead on the notion of
equality” (1990, 313).
Working within the framework of arguments supplied by Ronald Dworkin,
who is opposed to prohibitive measures, she tries to demonstrate that
egalitarian liberals such as Dworkin should support the prohibition of
pornography. She suggests that we have “reason to be concerned
about pornography, not because it is morally suspect, but because we
care about equality and the rights of women” (1990, 311).
Langton concludes that “women as a group have rights against the
producers and consumers of pornography, and thereby have rights that
are trumps against the policy of permitting pornography…the
permissive policy is in conflict with the principle of equal concern
and respect, and that women accordingly have rights against it”
(1990, 346). Because she is not basing her argument on the harm
principle, she does not have to show that women are harmed by
pornography. For the argument to be persuasive, however, one has to
accept that permitting pornography does mean that women are not
treated with equal concern and respect. It also seems that the
argument can be applied to non-pornagraphic material that portrays
women in a demeaning way that undermines their status as equals.
To argue the case above, one has to dilute one’s support for freedom
of expression in favor of other principles, such as equal respect for
all citizens. This is a sensible approach according to Stanley Fish.
He suggests that the task we face is not to arrive at hard and fast
principles that prioritise all speech. Instead, we have to find a
workable compromise that gives due weight to a variety of values.
Supporters of this view will remind us that when we are discussing
free speech, we are not dealing with it in isolation; what we are
doing is comparing free speech with some other good. We have to decide
whether it is better to place a higher value on speech than on the
value of privacy, security, equality, or the prevention of harm.
Fish suggests we need to find a balance in which “we must
consider in every case what is at stake and what are the risks and
gains of alternative courses of action” (1994, 111). Is speech
promoting or undermining our basic values? “If you don’t ask
this question, or some version of it, but just say that speech is
speech and that’s it, you are mystifying—presenting as an
arbitrary and untheorized fiat—a policy that will seem whimsical
or worse to those whose interests it harms or dismisses” (1994,
The task is not to come up with principles that always favors
expression, but rather, to decide what is good speech and what is bad
speech. A good policy “will not assume that the only relevant
sphere of action is the head and larynx of the individual
speaker” (Fish, 1994, 126). Is it more in keeping with the
values of a democratic society, in which every person is deemed equal,
to allow or prohibit speech that singles out specific individuals and
groups as less than equal? Fish’s answer is that, “it depends. I
am not saying that First Amendment principles are inherently bad (they
are inherently nothing), only that they are not always the appropriate
reference point for situations involving the production of
speech” (1994, 113). But, all things considered, “I am
persuaded that at the present moment, right now, the risk of not
attending to hate speech is greater than the risk that by regulating
it we will deprive ourselves of valuable voices and insights or slide
down the slippery slope towards tyranny. This is a judgement for which
I can offer reasons but no guarantees” (1994, 115).
These kinds of justification for prohibitions on hate speech suggest
that the permissive approach undermines free speech properly
understood. Even if hate speech or pornography does not cause harm (in
Mill’s sense) or offense, it has to be limited because it is
incompatible with democracy iteslf. The argument from democracy
contends that political speech is essential not only for the
legitimacy of the regime, but for providing an environment where
people can develop and exercise their goals, talents, and abilities.
If hate speech and pornography curtail the development of such
capacities in certain sections of the community, we have an argument,
based on reasons used to justify free speech, for prohibition.
According to Fish, the boundaries of free speech cannot be set in
stone by philosophical principles. It is the world of politics that
decides what we can and cannot say guided, but not hidebound, by the
world of abstract philosophy. Fish suggests that free speech is about
political victories and defeats. The very guidelines for marking off
protected from unprotected speech are the result of this battle rather
than truths in their own right: “No such thing as free
(nonideologically constrained) speech; no such thing as a public forum
purged of ideological pressures of exclusion” (Fish, 1994, 116).
Speech always takes place in an environment of convictions,
assumptions, and perceptions i.e., within the confines of a structured
world. The thing to do, according to Fish, is get out there and argue
for one’s position.
We should ask three questions according to Fish: “[g]iven that
it is speech, what does it do, do we want it to be done, and is more
to be gained or lost by moving to curtail it?” (1994, 127). He
suggests that the answers we arrive at will vary according to the
context. Free speech will be more limited in the military, where the
underlying value is hierarchy and authority, than it will be at a
university where one of the main values is the expression of ideas.
Even on campus, there will be different levels of appropriate speech.
Spouting off at the fountain in the centre of campus should be less
regulated than what a professor can say during a lecture. It might
well be acceptable for me to spend an hour of my time explaining to
passers-by why Manchester United is a great football team but it would
be completely inappropriate (and open to censure) to do the same thing
when I am supposed to be giving a lecture on Thomas Hobbes. A campus
is not simply a “free speech forum but a workplace where people
have contractual obligations, assigned duties, pedagogical and
administrative responsibilities” (1994,129). Almost all places
in which we interact are governed by underlying values and speech will
have to fit in with these ideals: “[r]egulation of free speech
is a defining feature of everyday life” (Fish, 1994,129).
Thinking of speech in this way removes a lot of its mystique. Whether
we should ban hate speech is another problem, albeit more serious,
similar to whether we should allow university professors to talk about
football in lectures.
Although Stanley Fish takes some of the mystique away from the value
of speech, he still thinks of limitations largely in terms of other-
regarding consequences. There are arguments, however, that suggest
speech can be limited to prevent harm being done to the speaker. The
argument here is that the agent might not have a full grasp of the
consequences of her actions(whether it be speech or some other form of
behavior) and hence can be prevented from engaging in the act.
Arguments used in the Skokie case would fit into this category and
there is evidence to suggest that watching pornography can cause
psychological damage the viewer . Most liberals are wary of such
arguments because they take us into the realm of paternalistic
intervention where it is assumed that the state knows better than the
individual what is in his or her best interests.
Mill, for example, is an opponent of paternalism generally, but he
does believe there are certain instances when intervention is
warranted. He suggests that if a public official is certain that a
bridge will collapse, he can prevent a person crossing. If, however,
there is only a danger that it will collapse the public can be warned
but not coerced from crossing. The decision here seems to depend on
the likelihood of personal injury; the more certain injury becomes,
the more legitimate the intervention. Prohibiting freedom of speech on
these grounds is very questionable for liberals in all but extreme
cases (it was not persuasive in the Skokie case) because it is very
rare that speech would produce such a clear danger to the
We have examined some of the options regarding limitations on free
speech and one cannot be classed as a liberal if one is willing to
stray much further into the arena of state intervention than already
discussed. Liberals tend to be united in opposing paternalistic and
moralistic justifications for limiting free expression. They hold a
strong presumption in favor of individual liberty because, it is
argued, this is the only way that the autonomy of the individual can
be respected. Feinberg suggests that to prohibit speech for reasons
other than those already mentioned means: “[i]t can be morally
legitimate for the state, by means of the criminal law, to prohibit
certain types of action that cause neither harm nor offense to any
one, on the grounds that such actions constitute or cause evils of
other kinds” (1985, 3). Acts can be “evil” if they are
dangerous to a traditional way of life, because they are immoral, or
because they hinder the perfectibility of the human race. Many
arguments against pornography take the form that such material is
wrong because of the moral harm it does to the consumer. Liberals
oppose such views because they are not impressed by states trying to
mold the moral character of citizens.
We began this examination of free speech with the harm principle; let
us end with it. The principle suggests that we need to distinguish
between legal sanction and social disapprobation as means of limiting
speech. As already noted, the latter does not ban speech but it makes
it more uncomfortable to utter unpopular statements. Mill does not
seem to support the imposition of legal penalties unless they are
sanctioned by the harm principle. As one would expect, he also seems
to be worried by the use of social pressure as a means of limiting
speech. Chapter III of On Liberty is an incredible assault on
social censorship, expressed through the tyranny of the majority,
because he claims it produces stunted, pinched, hidebound and withered
individuals: “everyone lives as under the eye of a hostile and
dreaded censorship…[i]t does not occur to them to have any
inclination except what is customary” (1978, 58). He
the general tendency of things throughout the world is to render
mediocrity the ascendant power among mankind…at present
individuals are lost in the crowd…the only power deserving the
name is that of masses…[i]t does seem, however, that when the
opinions of masses of merely average men are everywhere become or
becoming the dominant power, the counterpoise and corrective to that
tendency would be the more and more pronounced individuality of those
who stand on the higher eminences of thought. (1978, 63–4)
With these comments, and many others, Mill demonstrates his distaste
of the apathetic, fickle, tedious, frightened and dangerous majority.
It is quite a surprise, therefore, to find that he also seems to
embrace a fairly encompassing offense principle when the sanction does
involve social disapprobation:
Again, there are many acts which, being directly injurious
only to the agents themselves, ought not to be legally interdicted,
but which, if done publicly, are a violation of good manners and,
coming thus within the category of offenses against others,
may rightly be prohibited. (1978, 97 author’s emphasis)
Similarly, he states that “The liberty of the individual must be
thus far limited; he must not make himself a nuisance” (1978,
53). In the latter parts of On Liberty Mill also suggests
that distasteful people can be held in contempt, that we can avoid
them(as long as we do not parade it), that we can warn others about
them, and that we can persuade, cajole and remonstrate with those we
deem offensive. These actions are legitimate as the free expression of
anyone who happens to be offended as long as they are done as a
spontaneous response to the person’s faults and not as a form of
But those who exhibit cruelty, malice, envy, insincerity, resentment
and crass egoism are open to the greater sanction of disapprobation as
a form of punishment, because these faults are wicked and
other-regarding. It may be true that these faults have an impact on
others, but it is difficult to see how acting according to malice,envy
or resentment necessarily violates the rights of others. The
only way that Mill can make such claims is to incorporate an offense
principle and hence give up on the harm principle as the only
legitimate grounds for interference with behavior. Overall, Mill’s
arguments about ostracism and disapprobation seem to provide little
protection for the individual who may have spoken in a non-harmful
manner but who has nevertheless offended the sensibilities of the
Hence we see that one of the great defenders of the harm principle
seems to shy away from it at certain crucial points; even Mill was
unable to mount a defense of free speech on this “one simple
principle” alone. It does, however, remain a crucial part of the
liberal defense of individual freedom.
Liberals tend to justify freedom generally, and free speech in
particular, for a variety of reasons. According to Mill, free speech
fosters authenticity, genius, creativity, individuality and human
flourishing. He tells us that if we ban speech the silenced opinion
may be true, or contain a portion of the truth, and that unchallenged
opinions become mere prejudices and dead dogmas that are inherited
rather than adopted. These are empirical claims that require evidence.
Is it likely that we enhance the cause of truth by allowing hate
speech or violent and degrading forms of pornography? It is worth
pondering the relationship between speech and truth. If we had a graph
where one axis is truth and the other is free speech, would we get one
extra unit of truth for every extra unit of free speech? How can such
a thing even be measured? It is certainly questionable whether
arguments degenerate into prejudice if they are not constantly
challenged. Devil’s advocates are often tedious rather than useful
interlocutors. Sometimes supporters of free speech, like its
detractors, have a tendency to make assertions without providing
compelling evidence to back them up. None of this is meant to suggest
that free speech is not vitally important: this is, in fact, precisely
the reason we need to find arguments in its favour. But regardless of
how good these arguments are, some limits will have to be placed on
We have found that the harm principle provides reasons
for limiting free speech when doing so prevents direct harm to rights.
This means that very few speech acts should be prohibited. It might be
possible to broaden the scope of this principle, as Waldron attempts
to do, to include things other than harmful rights violations.
Feinberg’s version of the offense principle has a wider reach than the
harm principle, but it still recommends very limited intervention in
the realm of free speech. All forms of speech that are found to be
offensive but easily avoidable should go unpunished. This means that
much pornography and hate speech will escape censure.
If these arguments are acceptable, it seems reasonable to extend them
to other forms of behavior. Public nudity, for example, does not cause
serious harm and if it does offend some people it is at most a bit
embarrassing, and is avoided by averting one’s eyes. The same goes
with nudity, sex, and coarse language on television. Turning off the
television provides instant relief from the offense. Neither the harm
or the offense principles as outlined by Mill and Feinberg support
criminalizing most drug use, nor the enforcement of seat belts, crash
helmets and the like.
Some argue that speech can be limited for the sake of other liberal
values, particularly the concern for democratic equality. This
argument, unlike those based on harm and offense, has the potential to
allow significant limits on pornography and hate speech. The claim is
not that speech should always lose out when it clashes with equality,
but it certainly should not be automatically privileged. To extend
prohibitions on speech and other actions beyond this point requires an
argument for a form of legal paternalism that suggests the state can
decide what is acceptable for the safety and moral instruction of
citizens, even if it means limiting actions that do not cause harm or
unavoidable offense and which do not undermine democratic equality.
It has certainly been the practice of most societies, even
liberal-democratic ones, to impose some paternalistic restrictions on
behavior and to limit speech that causes avoidable offense. Hence the
freedom of expression supported by the harm principle as outlined in
Chapter One of On Liberty and by Feinberg’s offense principle
has yet to be realised. It is up to the reader to decide if such a
society is an appealing possibility.