Speech, Harm, and Mill
I think the current debates about free speech are a good thing, because for far too long there has been much less debate about free speech than a free speech regime urges us to have about everything else.
That’s Frederick Schauer, professor of law at the University of Virginia, in a recent interview with Richard Marshall at 3:16 AM. Near the beginning of the interview, the two discuss John Stuart Mill’s arguments about freedom of speech and thought in On Liberty.Professor Schauer first describes what he takes to be the typical interpretation of Mill’s arguments:
The most common reading of On Liberty takes Chapter 2 as an instantiation of Chapter 1. That is, in Chapter 1, Mill, famously, argues that it is illegitimate for the state to restrict personal conduct except to prevent harm to others. Preventing harmless immoralities, or preventing harmful conduct harming only the actor, are, Mill argues, beyond the state’s proper powers. And so when in Chapter 2 Mill argues for the “liberty of thought and discussion,” it seems straightforward to take thinking and discussing as coming within the Chapter 1 argument and thus as harmless. And this reading is reinforced by the childhood adage “Sticks and stones may break my bones, but names will never hurt me,” by the common civil libertarian acceptance of the harmlessness of speech, and by the frequency with which arguments from harmlessness are conjoined with distinct arguments from free speech, as is especially common (and often justified) in the case of arguments against restricting certain forms of sexually explicit speech whose only regulable attribute is its challenge to a certain form of moral sensibility about sexuality.
Professor Schauer thinks there’s a better interpretation available:
In contrast to the conventional reading I have just described, I think a better reading (whether it was Mill’s intended reading I will leave to others to decide) is that Chapter 2 is an argument not for protecting thought and discussion because they are harmless, but despite the harm they may cause. Thus, when Mill argues for the epistemic advantages of a regime of liberty of thought and discussion, I read him—or reread him—as saying that protecting even harmful thought and discussion is necessary to secure the epistemic advantages of a free speech regime, or to secure the advantages of intellectual character that come from requiring people to confront what they think false, and thus to understand why what they think false is actually false. And thus the conventional reading, which takes freedom of speech to be protected because of speech’s harmlessness, cannot explain (or justify) the protection of harmful speech, and thus offers a defense only of a weak and very limited free speech principle. My preferred reading explains and justifies a more robust free speech principle, although the question then remains whether such a robust principle is a good thing.
I think the current debates about free speech are a good thing, because for far too long there has been much less debate about free speech than a free speech regime urges us to have about everything else…
With respect to pornography, many of these debates have been frustrating because of confusions about labels. Traditionally, those who would restrict sexually explicit depictions have objected to the sexual explicitness, and have defined “pornography,” and the legal term “obscenity,” in terms of sexual explicitness. But although most sexually explicit literature until recently was distressingly celebratory of sexual violence and sexual coercion, there is nothing about sexual explicitness itself that needs to include such themes. Conversely, there is nothing about the verbal or pictorial endorsement of sexual violence, coercion, and exploitation that requires explicitness. So if we take Mill as arguing only for the protection of harmless speech, he turns out to have little to say about the regulation of speech that encourages or endorses or tolerates sexual violence, at least if we accept—and this is an empirical social science question and not a purely conceptual one—that widespread endorsements of sexual violence have a probabilistically causal relationship to the societal level of sexual violence. But if we take Chapter 2 as arguing for the desirability of protecting false views—the view that women who say “no” mean “yes,” to take one of the most notorious examples—then Mill provides support for the existing American legal and constitutional view, although, again, it is hardly clear that that that view should be accepted.
Much the same applies to many of the parallel debates on university campuses, many of which are focused on various forms of so-called hate speech, including race-based insults, vilification, and denigration. Again, there are important debates to be had about whether such speech ought to be protected in general, or ought to be especially protected (or especially unprotected) in university environments. But such debates will remain shallow as long as proponents of restriction insist that everything that is harmful ought to be restricted while opponents insist with equal vigor that the speech is harmless. Only by accepting that some of such speech is genuinely harmful can we have a fruitful discussion of when (if ever) and why harmful speech ought to be restricted, and when (if ever) and why harmful speech ought to be protected or tolerated.
You can read the whole interview here.
I think that Professor Schauer is absolutely correct that Mill thinks speech can be harmful. I think he’s also correct that, for Mill, the freedom of thought and discussion is a pre-condition for knowledge (at least, when it comes to matters in ethics, politics, and religion). However, I don’t think that Mill was only or even primarily concerned with the /value/ of knowledge. As I read him, his argument is not: without the freedom of opinion and discussion, we’d lack knowledge that our ethical, political, and religious beliefs are true. But this knowledge is valuable, so the freedom of opinion and discussion is valuable. This argument would not establish an absolute interdiction on proscribing heretical views, since the value of knowledge may surely be at times outweighed by the many other values at stake.
As I read him, Mill is focused on knowledge because he accepts a knowledge norm for action. If you don’t know that the heresy is a pernicious opinion, then you can’t forbid the expression of that heresy on the grounds of its perniciousness. The liberty to dissent is needed for you to know that the heresy is pernicious. So, if you take away the liberty to dissent, you’ve taken away the grounds of your knowledge that the heresy is pernicious. So you’ve taken away the grounds you had for prohibiting the expression of the heresy in the first place.
As I read Mill, /that/ is the argument for an absolute free speech principle—an absolute right to express and discuss any opinion whatsoever. The argument is that we would be justified in punishing somebody for the expression of an opinion only if we knew that that opinion is false. But, by punishing people who express that opinion, we undermine a necessary condition for our knowledge that the opinion is false. So such a law could never be just. Mill telescopes this argument when he writes:
“Complete liberty of contradicting and disproving our opinion, is the very condition which justifies us in assuming its truth for purposes of action”
He doesn’t see the knowledge norm for action as the controversial premise. Instead, he thinks he must persuade the reader that, without the liberty of opinion and discussion, you can’t know that your ethical, political, and religious beliefs are true. So he spends the vast majority of the essay defending that premise. But, as I read him, establishing that the possibility of dissent is needed for knowledge isn’t the primary goal of the essay. His primary goal is to argue that laws against dissent are always unjust. And he thinks this follows from the conclusion that the liberty to dissent is a necessary condition on knowledge, given the knowledge norm of action.Report
If we want a serious debate here, I think we should begin by stipulating some definition of harm, because the concept of harm, if you can even call something so incoherent a concept, is a mess.Report
“Harm” isn’t where the problem is, I’d say. “harm” = “a set-back to interest”. It is sometimes hard to know if people have suffered a real set-back to their interests, or if the actual set-back is as significant (or sometimes insignificant) as is claimed, but the basic idea, and core cases, are pretty clear. The trouble comes in trying to decide which sorts of harms are wrongful, or should be regulated for other reasons. (People also get into lots of confusion by having a moralized notion of harm – that is, saying that a set-back to interest isn’t a harm unless it’s also a wrong, but that’s just confused, and leads to needless trouble.)Report
We could stipulate that any set-back to an interest is a harm, but you would need to clarify what counts as a set-back. If “set-back” is understood broadly, then if I give my friend a box of chocolates rather than giving you one, your interest in eating chocolates (assuming for the sake of argument that you have one) is set back. But if harm is being defined that broadly, then Mill’s harm principle becomes ridiculous. Is this the most useful definition of harm for the debate in question?Report
To some degree all definitions here will be “stipulative”, so I’m not worried about that. But, I don’t think what you suggest is a problem. I think we can, and should, take any set-back to interest to be a harm, and that this is a plus. Take your example: if you could give me or someone else some chocolates, and give them to the other person, and this prevents me from having some, and I had an interest (taken purely descriptively) in having some, why not say this is a harm? It’s likely a trivial harm, but it’s good to be able to say that. (It might not even be an all-things-considered harm – I might have an interest in eating chocolates because I like them, and an interest in not doing so because of my health, and the later might out-weigh the former, making this a harm when considered narrowly, but not all-things-considered. This account of harm makes it easy to say this, and to see what’s happening. That’s a plus.)
This definition also helps with the sorts of worries that Preston is interested in below, because it forces us to think about what we mean when we say that some speech is “harmful”. We have to ask, “whose interests are set back by this type of speech, and how?” We then also have to ask “How seriously are these interest set back”, and “do the people have a right to not have their interests set back in this way/case/situation?” And, we see better than on moralized accounts of harm that these are all three distinct questions. That’s the first step towards clarity here.
This account also lets us have a clear idea of wrongless harms and harmless wrongs, which moralized accounts can’t easily deal with. That’s a problem, as those are useful and intuitive categories. So, I see no real disadvantage here, and lots of strengths. A definition won’t make everything easy – especially not largely empirical issues about how much harm someone has received in particular cases. But, it’s wrong to think a definition could or should do that.Report
That’s fine. I was arguing in favor of using a stipulative definition, not against doing so, and I also agree with you that your definition is better in many ways than moralized definitions of harm. Of course, on your definition, all speech harms someone, and so the relevant debate simply shifts to moral questions about when speech infringes or violates rights, or when it unjustifiably sets back interests, which is where it should have been all along.Report
“[I]f you could give me or someone else some chocolates, and give them to the other person, and this prevents me from having some, and I had an interest (taken purely descriptively) in having some, why not say this is a harm?”
Because it flies in the face of any common usage of the concept?
And that it’s trivial isn’t much the point. Instead of chocolates, it could just as well have been the life-saving antibiotics I gave to my gran, instead of some random Bloggs I’ve never met. But, there’s no sense of the word “harm” that I recognise by which I harmed Bloggs by not giving him my antibiotics.
Moreover, we’ve already got a good concept for this sense of “harm.” It’s the one consequentialists are, more-or-less, interested in—namely, X is worse off than they would otherwise have been.
The morally interesting sense of harm is one that is pro tanto grounds for wronging. (And isn’t buying into that sense of harm and the wrongings it gives rise to, one of the paradigm markers of non-consequentialism?)
If the fact that A harmed B isn’t even pro tanto grounds for A to have wronged B (/for B to have a complaint against A)—and it isn’t in the case I quoted— then we’re not talking about any sense of harm that I recognise (either on the street or as an ethicist).
(And, note that this doesn’t get into the confusion alluded to above: “People also get into lots of confusion by having a moralized notion of harm – that is, saying that a set-back to interest isn’t a harm unless it’s also a wrong.” Since A can wrong B without acting wrongly.)Report
I’m going to second kailadraper’s suggestion, and try to expand on it. There’s a strong case to be made that the debate about free speech needs to be more explicitly focused on just what “harm” is supposed to mean.
Schauer gives us examples about sexual violence in pornography, and he then refers to “parallel debates” on University campuses about racism. He concludes “such debates will remain shallow as long as proponents of restriction insist that everything that is harmful ought to be restricted while opponents insist with equal vigor that the speech is harmless”. But the move from sexual violence to debates about racism on college campuses, as a basis for inferring that the way to deepen the debate on free speech restrictions lies in considering whether harmful speech should be permitted in spite of its harm, overlooks two things: one, free-speech advocates often take issue with the broadened definition of ‘racism’ that is sometimes used, so there’s no agreement about just what counts as racism in the first place; and two, terms like ‘harm’ are also undergoing conceptual bloat–my favorite example came from the fallout from Rebecca Tuvel’s article about transracialism in Hypatia, where one philosopher claimed that the paper contained “egregious levels of liberal white ignorance and discursive transmisogynistic violence”.
Schauer may have a fine reading on how to think about Mill’s stance on balancing considerations about harm and the value of free speech. But given that there’s no agreed use of these terms across the debate, with oppontents to free speech restrictions taking issue with the notions of ‘harm’ and ‘violence’ that advocates of those restrictions sometimes employ, we can’t raise the question of whether harmful speech should sometimes be protected without first settling what we mean by “harm”.Report
I agree that relevant concepts such as “harm” are ambiguous, and that this matters, at least for purposes of conceptual clarity when teasing out when speech crosses lines allegedly drawn by Mill.
But such debates are arguably mooted by the reality that, once people are given the power to define “harm,” and to further judge the indirect consequences of that harm as part of a broader calculus of deciding which speech should be protected, we’ve already gone too far. Because obvious aspects of human nature, combined with historical experience, teaches that empowering some to make this decision leads to huge problems. For example, indeterminacy inheres in any conception of harm, regardless of how thoroughly the concept is explicated in the abstract. Nature hates a vacuum, and so the conceptual vacuum created by indeterminacy gets filled with political and ideological agendas, as well as with the capricious intuitions of deciders reacting to specific speech in specific cases.
In other words, the problem might be the assumptions that people entertain about our ability to employ line-drawing methodologies in the first place. I realize that, as a local philosophical exercise, this isn’t a problem; but the exercise is taking place, presumably, as part of broader effort to solve real problems.Report