What is lynching?

A blog about the violent punishment of offences by regular people

Source: Library of Congress

CONTENT WARNINGS. Description of violence. Reference to, but no description of, sexual violence.

So, what is lynching?

Let’s begin with Regina Bateson’s definition of vigilantism, this being ‘the extralegal prevention, investigation, or punishment of offences.’

I choose to start with Bateson’s definition, partly because she wrote a whole article to justify her choices¹, which none of the other (many, many) definitions of lynching do, but also because it introduces some useful concepts. First is ‘extralegality’, which is used as it ‘implies an action that is moving in the same direction as the law, but exceeding its scope or severity.’ Second is ‘offences’, which are violations of rules of behaviour that include, but are not necessarily, crimes. Both concepts are useful, because they mean we should see vigilantism as existing in a continuum with other types of policing, the difference being the actions of vigilantes are not formally sanctioned by codified state authority. Bateson is very clear that cops can also be vigilantes.

However we’re not here to look at all vigilantism, but lynching in particular, which I’m using as a shorthand to hold two features of vigilantism constant. First, lynching is primarily about punishment. The punishment is often justified as preventative, and there is an investigative component to it, but the top priority is often simply to repay a hurt in kind. Second, lynching is a civilian activity. It is not the work of ‘gangs, rebels and regular security forces’. It’s this that makes it interesting.

Interesting?

Yes, right. The thing is, when people in the US, Europe and other countries where the work of creating order is done almost exclusively by organisations created and funded by the state, think about policing, it’s those organisations that dominate our (statist) thinking. Policing becomes something that only the police do, and as such, when we statists want to improve or critique policing, we think about this in terms of improving or changing the organisation. Lynching however disrupts this. It demonstrates that policing is not just an act designed by bureaucrats to serve a broader social or ideological mission. It is also something people just do.

Is it not something only racist Americans do?

The most famous lynchings are definitely the ones performed by racist Americans in the ex-slave states, mostly between 1880 and 1920, but with a significant trail into the 1930s/40s. As a result of their pre-existing fame, I chose not to read much about them. But it is worth touching on their extent, and their cultural context.

For the first, the latest data collection effort, conducted by Charles Seguin and David Rigby, building on earlier work by Tolnay & Beck, identifies 4,467 victims of lynchings in the USA between 1883 and 1941.² Of these 88% were killed in former slave states. Of the 3,265 black victims of lynching in this period, 97% were killed in the former slave states. Focusing only on those killed in lynchings though, as Seguin and Rigby do (they define the events they study as ‘extrajudicial killing’) likely underestimates the extent of the practice. The lethality rate of lynchings in other places and periods was not 100% (as we will see), so we shouldn’t assume that was the case here. This suggests a currently-invisible universe of non-lethal but still-terrifying acts of corporal or attempted punishment.

The other feature of this Southern lynching is how closely linked it seems to be to the practices of slavery. This is not something I’ve reviewed personally, but Seguin and Rigby review prior work that find both that the practices of lynching were inherited from punishments deployed by slavers, and, that lethal lynchings were more common in US counties with cotton-planting. This implies something key. Lynching in the US South was an inherently racist practice, but this is because in this place it was a legacy of racist practices. Lynching may not be the work of organisations, but that is not to say it is disorganised - rather that the organising factor is culture. This is particularly evident in another time and place of the USA: the California Gold Rush.

How is lynching organised?

In the gold rush mining settlements of California (1840s-50s), lynchings were often performed through ‘trials’ that mirrored the formal legal procedures of the Western USA.

After the offence occurred, a crowd would either spontaneously gather, or be purposefully collected, with the purpose of trying the suspect. Once assembled the crowd would adopt a formal meeting structure, and elect a president or foreman to officiate. They would then select a judge and a jury by vote, and sometimes with the suspect’s participation. In some instances they would even appoint lawyers for defence and prosecution.

After hearing evidence, the lynch jury would deliberate and decide both verdict and punishment. The former was usually (but not always) guilty. Punishments included whipping and branding, but more serious crimes like murder tended to result in hanging. The miners doing the lynching of course did not have gallows. Instead, they would set the condemned on a wagon (or other platform), tie a rope around their neck, drape it over a tree limb, and then drive the wagon out from under the condemned’s feet. This person would then be strangled to death.³

None of this procedure was adopted as a result of conscious organisation. Andrea McDowell, whose book We The Miners⁴ is the first in depth treatment of the lynch trials, takes pains to distinguish them from what were called at the time ‘vigilantes’, e.g. groups with formal constitutions and defined memberships who were also engaged in extralegal policing. Nor were these procedures imposed by a state. There was no government in California when the lynch trials began, and when there started to be a government (with state courts and state police) the lynch trial procedure broke down, as the miners tried to enact their punishment as quickly as possible before the state’s agents could intervene. People were following a script no-one needed to write.

The lynch trials of the California Gold Rush are not common practice. In most other lynching tales I’ve read the judgement that a person needs to be punished emerges without even this minimal deliberation. It just emerges, suddenly. The following lynching that occurred in South Africa in 2006, related in Nicholas Rush Smith’s Contradictions of Democracy, feels emblematic:

The victim, upon seeing the men in the squad car, loudly screamed they were the rapists. Alert neighbors rushed from their homes and surrounded the car, screaming at the suspects and policemen. As the crowd grew, people yelled for the suspects’ blood and demanded the policemen hand them over. The crowd began to throw stones and the panicky police, having sustained injuries, ran. As the officers fled, the crowd ripped the car open and dragged out the suspects. A hail of stones, rocks, and bricks felled the men.⁵

But we should not read this immediacy as a lack of organisation. Desire has patterns too, features that may be found across multiple brains. Next section attempts to discern this pattern using data from modern-day Latin America.

Who do people want to lynch?

This is a question researchers have attempted to answer, using two methodologies: survey conjoint experiments, and, descriptive analysis of a new database of lynchings.

A conjoint experiment is a survey technique in which an interviewer shows the person answering the survey two descriptions of a thing, and asks this respondent which they prefer, or think better fits some condition. This process is then repeated, with the interviewer showing the respondent multiple description pairs, each time varying random features of the descriptions. Do this with enough respondents, and you end up with data that shows which elements of a description most contribute to it being selected.

Freire & Skarbek (2020) and Dow et al. (2023)⁶ use this method to figure out what types of crime, and what types of criminal, people in Brazil and Guatemala respectively think it is more justifiable to punish extrajudicially. It should be noted that effect sizes found in the Guatemala study are tiny, while in F&S they’re much larger, even when the conjoints are testing similar crime/criminal features, so either something about the projects’ different methodologies, or a difference between Brazil and Guatemala, or both, means how impactful these features are is not geographically consistent. That said, the figures that do have the most impact in each context are consistent.

Each study finds that of the different crime/criminal features tested, the features with the biggest impact tend to be the gender of the perpetrator, and the type of crime committed. In both countries people report thinking that extrajudicial punishment is more valid when men do it, and when the crime is more serious (e.g. murder or rape, or a crime that targets a child). Less serious crimes and crimes committed by women are less likely to be seen as valid for lynching. The Guatemala study finds an explanation for the source of this pattern. ‘We find that our results look similar if we focus on which of the two crimes respondents think should be punished more severely, rather than which should be subject to vigilantism.’ The desire to punish seemingly stems from a need to respond with appropriate severity.

But, at the end of the day, these conjoint experiments are a study of what people who have not necessarily lynched anyone claim to believe about the legitimacy of extrajudicial punishment. Do real-world lynchings match up?

Nussio & Clayton (2024) provide us with an opportunity to answer this.⁷ They have assembled a dataset of 2,818 lynchings reported in news articles covering Latin America and published between 2010 and 2019. Unlike Seguin & Rigby’s US lynching database mentioned above, N&C state that the events in their database do not have to have a fatal outcome, a choice that makes their data more valuable for getting a sense of the practice in its entirety (within the constraints imposed by the necessary reliance on newspaper reporting, which obviously will not record every lynching). So, who were the targets of these lynchings?

Given the limited time available to collect data, information on the age and gender of lynching targets in the N&C database is only available for four countries: Brazil, Mexico, Colombia and Guatemala. These show that men are overwhelmingly more likely to be lynched than women (91% vs 9%), and younger people (18-35s) are more likely to be lynched than older. In terms of men though, this is less likely to be the proof of preference than it is the fact that men tend to be far more likely than women to commit crimes.⁸ So while N&C data is in line with reported preferences, the story here is muddled by the demographic disparity of crime.

You could make a similar claim about commonality when it comes to what offences are likely to trigger a lynching. Counter to the results of the conjoint, in most lynchings the suspected offence is most often theft (1,745 examples, 62%). Suspicion of of murder (439 cases) and child abuse (359) come a distant second and third. Theft is a far more common crime than murder or child abuse, so potentially that is why reported lynchings happen to people that our survey respondents tend to consider less legitimate targets of extrajudicial violence. However, I am not immediately comfortable accepting this. I think the discrepancy here may be more fundamental.

The conjoint experiments are expressly comparative. They generate findings by comparing fictional crimes. But it seems unlikely that when a person decides to do a lynching, they might dissuade themselves by contemplating how bad the crime they suspect has happened is, relative to other potential crimes. What they will likely be acting on instead is, I would argue, a volume of emotion inspired in them by the perceived crime. People may be more likely to punish murders than thefts. But a theft can still take a person over the edge into violence, and ultimately, what we want to know is why that might happen.

What takes a person over the edge?

Ultimately, this is a question about emotions and the beliefs that undergird them, and at present there seems to be one study that examines these.

Enzo Nussio designed and organised an in-person survey in Mexico City in 2022.⁹ The survey had two key sets of questions. In the first set of questions the respondent was read an example of a locally-normal lynching of a thief, and then asked whether if such an event occurred they would stay and watch, or join in, if they had watched or participated in a similar lynching in the past, and what emotions the lynching vignette made them feel. The second set of questions focused on the respondents’ moral beliefs. These involve questions about what considerations are relevant for respondents determining right from wrong, and also agree/disagree questions on a slate of various ethically revealing topics.

Nussio’s trick is that for half of his respondents he put the set of moral questions ahead of the lynching questions in the survey, so that respondents were thinking about their own morality for 5-10 minutes before thinking about how they would react to a lynching. For the other half of respondents, the morality questions came after the questions on lynching. Nussio’s hypothesis is that, because harming others is the closest you can get to a universal moral evil, if a person is thinking about their own morality when facing a decision to enact collective violence, they will be less likely to do so.

Nussio’s experimental results support this hypothesis, with those primed to think about their own morality being a little less likely to say they would support the hypothetical lynching, and more likely to report feeling compassionate emotions towards the thief.

How does this relate to going over the edge? Well, the result here, singular and based in hypotheticals as it may be, offers an improvement to how I think about lynching. I said at the top of this essay that what makes lynching interesting is that it is something people just do. But no-one just does anything. Action follows thought, and people possess a lot of competing thoughts, some that might tip them over into policing people with brutality, and some that might keep them away from it. The reason why the wrong set of thoughts is in a person’s head at the wrong time is a complex one, requiring far more investigation than I have time for this month. But, I feel the shift in attention in myself, towards thinking about the source of violence as the inner workings of a mind not an organisation, to be productive. I began this blog asking what lynching is. The answer is it’s a very specific man-made horror, that with further examination, can maybe help me think more accurately about man-made horror in general.

1. Regina Bateson, ‘The Politics of Vigilantism’, published in Comparative Political Studies in 2020. It’s great!

2. Charles Seguin and David Rigby, ‘National Crimes: A New National Data Set of Lynchings in the United States, 1883 to 1941’, published in Socius in 2019.

3. This is bluntly horrible, but I want to make clear (as does McDowell) that despite their veneer of procedure, the California lynchings were acts of brutality.

4. Andrea McDowell, We the Miners: Self-Government in the California Gold Rush, published in 2022

5. Nicolas Rush Smith, Contradictions of Democracy, published in 2009, page 81. The punishment enacted by the crowd escalates from here, but my point does not depend on the grisly details.

6. Danilo Freire and David Skarbek, ‘Vigilantism and Institutions: Understanding Attitudes toward Lynching in Brazil’, published in Research & Politics in 2023; David A. Dow, Gabriella Levy, Diego Romero and Juan Fernando Tellez, ‘State Absence, Vengeance, and the Logic of Vigilantism in Guatemala’, to be published in Comparative Political Studies.

7. Enzo Nussio and Govinda Clayton, ‘Introducing the Lynching in Latin America (LYLA) dataset’, published in the Journal of Peace Research in 2024.

8. I do not off the top of my head know the right data to support this claim. However, Figure 2 in this UK report on prisons provides statistics on the gender balance of people being admitted to prison, and its maleness is stark. If anyone has other raw data that backs this claim, please comment and pass it along!

9. Enzo Nussio, ‘How Moral Beliefs Influence Collective Violence. Evidence From Lynching in Mexico’, published in Comparative Political Studies in 2023.