Social Signaling and the Undoing of the Harm Principle

By Nathaniel Counts

In On Liberty, John Stuart Mill asserted that “[t]he 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.”  This has since become known as the harm principle and is foundational for much of American political discourse, especially for libertarianism and civil rights.  At the time of Mills’ writing, On Liberty having been published in 1859, this logic could protect a lot of conduct that involved consenting adults or did not appear to directly impact others.  If the harm principle was controlling, we could never have a soda ban.  Today however, with our advances in social science, clear lines of harm and no harm have become fuzzy.

Few people buy cigarettes in a vacuum.  Someone offers you cigarettes or you see other people smoking and then you buy cigarettes.  This phenomenon, in which we make decisions based on the decisions of others in relationship to our esteem for them (if we see someone we respect smoking, we will be more likely to smoke; if we see someone we do not respect smoking, we will be less likely to smoke) is called social signaling (The Origin, Development, and Regulation of Norms is a great article on a related topic, the generation of norms, which explains this phenomenon quite well).

Social signaling is not inherently bad – we can receive signals to donate to charity, to be a better parent, to come to the aid of strangers – social signaling is neutral and dependent on the substance of the signals themselves.  Some of these signals will make us worse off than if we were allowed to make an independent decision.  It is hard to imagine many people who, having never encountered a smoker, would find cigarettes on the shelf in a store, figure out what they are for, and then buy them.  Such people very well may exist, but it would certainly be a lower proportion of the population than currently smokes today.  Because of this, many people may experience worse health outcomes caused, in part, by the actions of others.

This is in part why the non-smokers rights movement was so effective in reducing new smoker initiation (this study is a good example).  The non-smokers rights movement made it illegal to smoke in many public spaces, so the act of smoking was less often a signal to others.  Note that this does not militate for allowing smoking in private spaces – family members will still observe the smoking, these laws only limit the scope of the signaling.  For some health concerns like obesity, prohibiting consumption of unhealthy foods when anyone else is able to observe would be insufficient as the obesity itself may be a signal to others.

Thus, social signaling presents a problem for the harm principle – actions we thought only harmed ourselves actually do harm others and no decision is without externalities.  If we accept this, then the harm principle cannot be a basis for denying state regulation of certain activities that pose risks to public health.  The soda ban cannot be proscribed on these grounds.  To state this proposition in its starkest terms, based on the harm principle alone, the government could make it illegal for you to drink a soda at the movies because, in doing so, you are hurting the other patrons.



Nathaniel Counts was a Student Fellow during the 2013-2014 academic year. At the time, he was in his third year at Harvard Law School. He was interested in the role of law and lawyers in the treatment of mental health issues, with a focus on behavioral disorders, including intersections with the criminal justice system. He was also interested in the use of a right to health care in human rights lawyering and international development. Nathaniel graduated from Johns Hopkins with a major in biology and a minor in entrepreneurship and management. Prior to law school, he studied creative writing at Bar-Ilan University in Israel. His past research focused on the federal government’s response to marijuana legalization, including recommendations for public health initiatives; he published an article on this subject in the Gonzaga Law Review in 2014.

3 thoughts to “Social Signaling and the Undoing of the Harm Principle”

  1. Hi Nathaniel – A few questions for you: Don’t you think there is a difference between affecting someone else’s choices (often involuntarily) and harming them? What about the difference between acts and omissions? And personal responsibility for one’s own health behaviors? And limitations on duties to others – for example, do I really have a duty to the entire universe to not let them witness my bad eating habits, lest they imitate them?

    Where I thought you might be going with this is down the path of externalities – when we eat poorly and have poor health, others are affected by virtue of having to share the cost of care, lost days of work, etc.

  2. Thank you for the great questions! I will respond to them in order:

    1. It certainly feels different for an older kid to give a younger kid a cigarette and encourage her to smoke than for the younger kid to see the older kid smoking and decide to smoke, but if the result is the same, the kid starts smoking and faces the associated health risks, the child has suffers the same harm. The culpability for incurring the harm is distinguishable, but I think the harm principle would perhaps be too narrow if it only covered intended harms.

    2. Personal responsibility for health could well be decisive here even if the harm principle alone does not militate against regulation of smoking and eating behaviors. What might be controlling here is how big of an influence we believe social signalling has on behavior, and how much influence we are willing to tolerate before we are willing to say that an individual no longer has personal responsibility because the decision is essentially being made for her by others. It may also matter how large the harm is – even the slightest push toward a terrible outcome may be unacceptable.

    3. This extent of our duty to one another is something I feel must be determined by the community, i.e. the members of a community should agree as to how much duty that should owe to one another. If the decided upon level of duty allows individuals to signal harmful behaviors to one another, then it should not be regulated, but the harm principle on its own would not prevent the regulation.

    The externalities argument is an exciting one but worrisome to me. Although the economics supports regulation, I do not think it is what we are ultimately concerned with – we are actually concerned with the individual’s wellbeing as weighed against her autonomy. If the externalities are more a convenient justification, I worry that this type of reasoning may one day cut against an important public health measure and lead to disastrous consequences (for example, if EMTALA, which requires hospitals to treat anyone who goes to the emergency room, and Medicaid and Medicare were never passed, most public health measures could not be justified as cost-saving.).

  3. Another interesting piece Nathaniel!

    I agree with point #3 in your response comment above – sometimes there are only externalities because we have manufactured them through the design of our social service institutions.

    If I understand you, you are saying that because of the effect of signaling, by smoking (mostly in public) you are harming others by causing them to take up the harmful activity of smoking? That’s quite a stretch, and I wonder if the probability analysis of that occurring makes the harm (assuming we accept signaling as a harm) so remote as to be de minimis? At a minimum, the issue of remote causation arises. You cannot claim that public smoking CAUSES harm to others via signaling (no studies show causation).

    You should check out Joel Feinberg’s work, as he offers the most detailed analysis of the harm principle ever performed (albeit in the criminal law realm). He defines harm as a wrongful (rights-violating) setback to interests. Feinberg’s definition also requires something you glossed over in your reply comment – intent – A must intend to harm B or is reckless or negligent in the creation of the harm. Does the public smoker intend to signal to others which you allege is a harmful act? Not likely.

    My biggest critique would be that harm principle doesn’t even come into play in the signaling analysis. Here’s why: consent. The old liberal argument goes that you are not harming anyone if they have consented to the “act”. This gels with Feinberg’s definition, since you can’t violate anyone’s rights if they consent to what is done to them, nor can you setback their interests by playing an indirect causal role in their decision to smoke (a remote causation issue).

    Consider this example. I am an avid skier and encourage you to take up the sport (more direct involvement than in your signaling case) and then ten years later you get injured in a horrific accident. Surely you wouldn’t track me down and point to my getting you into skiing in the first place as causing that harm. I haven’t harmed you because you made the ultimate decision of whether or not to go skiing (an inherently risky activity). You harmed yourself. Same goes for signaling. The choice remains with the individual as to whether or not to take up smoking. They may want to because they are exposed to social signaling, family tradition, advertising, etc., but the final decision is still their own, and therefore they consent to any “harm” that may befall them by smoking so long as they are aware of the risks, there is no deception, and other usual caveats. Simply put, motivation is different from action.

    So to summarize: signaling is not harmful, on account of: (i) lack of intent (ii) remote causation, and (iii) consent vitiating the harm.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.