Originally written in September 2024. I’ve only made some light style edits before posting here.

Among the many problems to solve in the world, the existence of sadistic tendencies among human beings may be an important one. This post explores this phenomenon. My exact original motivation for writing it was that understanding how different forms of sadism evolved in humans might give us insights vis-à-vis how they might also evolve in TAIs and how we can prevent that. But improving our understanding of sadism can be useful in many other ways.

Main takeaways

• I show that (some) people have preferences that seem at least partly motivated by sadism (a genuine preference for there to be more suffering, at least under some conditions), and explain how these differ from merely “sadistic-looking preferences” and why this matters (see §2).
• I go through compelling explanations as to why different forms of sadism evolved in humans (see §3.1).
• I also discuss (potential) sociological risk factors for sadism (see §3.2).
• I propose research questions future research on the topic could address (see §4).

1. What I exactly mean by “sadism”

I here define sadism as an intrinsic preference for there to be more suffering (given a certain context). This includes both unconditional cases (i.e., no matter what the context is) and the ones conditional on, e.g., some belief that the suffering is “deserved” (tribalistic and/or retributivist sadism). This excludes “sadistic-looking” behaviors that aren’t actually well-explained by a terminal preference for suffering per se (see §2 for examples that help distinguish the two). This also excludes cases often associated with sadism, although meaningfully different, such as non-sadistic manifestations of the other Dark Tetrad traits (narcissism, machiavellianism, and psychopathy) and of antisocial personalities.

Such a definition is fairly different from what people usually have in mind when they discuss sadism – see e.g. this Wikipedia page and Foulkes (2019), section 2, which discusses different academic definitions.

In particular, my definition doesn’t solely focus on cases where the sadist derives pleasure from the suffering of some being(s). It also includes those where the satisfied preference for there to be more suffering doesn’t necessarily make the sadist experience hedonic value, at least not in the way it is most commonly defined and/or interpreted. For example, someone might deeply want someone else to suffer thinking “they deserve it” (for whatever reason), try to make that more likely, and find it fair if it happens without necessarily taking pleasure in knowing they suffer. Although it may be very much worth differentiating these two different types of phenomena on a psychological level for practical reasons, I believe it makes sense to use sadism as an umbrella term including both.

I landed on this definition to focus on what seems most concerning to me: a direct, although sometimes contextual, preference for there to be more suffering.

2. What are some clear manifestations of sadism in humans?

It can easily be argued that many sadistic-looking behaviors (some) humans show are not necessarily indicative of an intrinsic preference for someone to suffer (at least given a certain context). Consider bullfighting, which, in the eyes of many animal advocates, is a sadistic form of entertainment. While taking enjoyment from seeing a bull being forced to repeatedly charge someone unsuccessfully before being stabbed to death is very much sadistic-looking, this could quite well be explained by, e.g., “mere” lack of empathy towards the bull paired with the enjoyment of watching something impressive.^[1] This observation generalizes to some of the criteria used within the Short Dark Tetrad scale (SD4) and even within the Short Sadistic Impulse Scale (SSIS)^[2] to measure sadism (e.g., the criteria “Watching a fist-fight excites me” part of the sadism sub-scale of the SD4 and “I have humiliated others to keep them in line” part of SSIS). These don’t seem appropriate to directly measure actual sadism rather than a propensity to enjoy and/or endorse sadistic-looking acts (plausibly for non-sadistic reasons).

The present section, however, aims at compiling actual evidence of actual sadism in humans, the overall key thing here being that some people self-report and/or demonstrate having (through their acts) preferences that can hardly be explained without sadism being at least one of the drivers:

• In various surveys,^[3] a substantial number of respondents “agreed” with statements that seem to be pretty solid indicators of sadism (“Hurting people would be exciting”, “I have hurt people because I could”, “I have hurt people for my own enjoyment”, “I would enjoy hurting someone physically, sexually or emotionally”, “I enjoy seeing people hurt”, “I have fantasies which involve hurting people”. “Sometimes I get so angry I want to hurt people”. “People would enjoy hurting others if they gave it a go”).

• There are plenty of stories of extreme physical torture where it seems pretty reasonable to assume the perpetrators intrinsically wanted – among other things – their victims to suffer.

  • 20th-21st century criminal cases where the oppressor keeps coming up with new and elaborate ways to torture their victims that stick out as lacking motives other than pure sadism.^[4] They sometimes target total strangers they have absolutely no link to. They sometimes torture their victims knowing they will kill them (and/or themself) shortly after.

  • Public torture (with horrifyingly complex methods) was common in e.g. medieval Europe, Ancient Rome, and during the Spanish Inquisition. It is rarer in the 21st century but still occasionally happens in some countries. This is a tribal and/or retributive form of sadism (see Baumann 2022).
• Relatively elaborate forms of sadism that don’t involve extreme torture but still can hardly be dismissed include

  • Criminal acts of sheer vengeance (retributive sadism).^[5] One prominent example is revenge porn. There are also specific occurrences that appear more extreme such as Marvin Heemeyer’s bulldozer rampage after losing a zoning dispute.

  • What we can call spiteful sadism where the suffering the sadist wishes would occur is specifically perceived as coming from the frustration of some goal or principle the victim has.^[6] It can be, e.g, someone intentionally eating more meat and/or being more vocal about how much they like it just to enjoy how much this upsets a vegetarian.

• Some of the human acts that can be stamped everyday sadism (see Buckels et al. 2013, Paulhus and Dutton 2016), because of their relative commonality,^[7] also appear hard to explain without some form of pure sadism being at least partly at play, such as

  • Epilepsy trolling. “Some videos and GIFs shared on social media platforms have the potential to trigger seizures in people with photosensitive epilepsy. Unfortunately online trolls know this, and sometimes send targeted posts or messages to people with epilepsy online, in an attempt to trigger a seizure.” (Epilepsy Society n.d.)

  • Griefing (deliberately annoying other people in multiplayer video games for the sake of it) where the enjoyment seems to come at least partly from knowing and/or witnessing how frustrated the target is.

  • Sharing and seeking out videos of people being hurt or humiliated, without any clear instrumental purpose.

  • Non-public bullying behaviors that don’t affect anyone’s social status (e.g., between siblings at home).

  • Hazing.

  • (Retributive sadism examples)

    • Someone trying to ruin the life of their ex-partner.

    • Most cases of customer retaliation, where, e.g., the desire to leave invecting reviews seems at least partly motivated by an intrinsic need for retributivism.

3. Why does sadism exist?

This matters because we might have an easier time trying to solve the various problems posed by sadism if we understand what did/does allow it to exist.

  3.1 How did sadism evolve?^[8]

This section goes through intuitive compelling hypotheses (that are not meant as mutually exclusive):

(i) Retributive sadism might generally be due to the adaptiveness of punishing those who wronged you, for obvious game-theoretic reasons (see e.g. Macé et al. 2023, Appendix). One way to make sure a given agent doesn’t let others exploit them is to make this agent develop an intrinsic desire not only to execute such punishments but also to witness the effectiveness of their punishment on the target.

The following quote from Jackson et al. (2019) and the references therein support (i):

[W]hen people are calculating the appropriate severity of retributive punishment, they are more attuned to whether the punishment matches the original transgression than whether it deters future harm (Carlsmith & Darley 2008). People playing economic games will also take revenge when they know they will not encounter their partner again, which would not make sense if revenge was solely intended to deter future harm (Fehr & Gachter 2000). Interestingly, vengeful people often feel less safe from future harm ¨ than do nonvengeful people (Akın & Akın 2016), indicating that people do not commonly take revenge because they think it will protect them. This evidence suggests that, of the many proximal predictors of revenge, deterrence may be among the least influential.

The next hypotheses focus on explanations that have nothing to do with retributivism or vengeance to try to identify explanations independent from (i), although there of course sometimes is a substantial overlap of explanations.

(ii) The kind of sadism that motivates public torture may reflect that wanting the enemy to suffer helps you beat them, intimidate them, survive, and spread.

Nell (2006) argues that sadism was helpful while hunting, which arguably backs some specific version of (ii).

(iii) Something very similar can be said for everyday sadism, except that the focus might be more on dominating those closer to you (not necessarily identified as “enemies”, at least not in the same sense as above) to ensure your survival and reproductive success.

Serotonin falls in scarcity (Crockett et al. 2008) which leads to sadistic behaviors (Crockett et al. 2013). This seems quite consistent with both (ii) and (iii). Also, Nicolas Macé et al. (2023) write:

Animosity towards outgroup members has been hypothesized to have been selected for in the context of zero-sum competition between groups for opportunities to mate with their respective female members ((Choi and Bowles 2007; Tratner and McDonald 2019); cf. (Konrad and Morath 2012))

(iv) 20th-21st century criminal cases of extreme torture are a bit more puzzling. While the last two examples treated above suggest that wanting others to suffer was a competitive feature in some situations, a substantial distributional shift is needed here to explain why some would be motivated to kidnap humans or non-human animals they don’t know and torture them secretly. This distributional shift however seems like a plausible theory for such relatively rare cases.

  3.2 Sociological risk factors for sadism

Let’s flesh out some intuitive hypothesized risk factors (that are not meant as mutually exclusive) before briefly looking at whether the literature proves them and to what extent:

(a) Repeated exposure to suffering and sadistic-looking acts in a way that habituates someone to – or sometimes even teaches them to enjoy – these, making their empathy less likely to be stronger than their potential sadistic impulses.

(b) Someone experiencing things that relatively reduce their empathy and/or altruism (e.g., things they perceive as deeply unfair, making them “lose faith in others”) and hence make them more likely to act on their potential sadistic impulses.

(c) Factors that develop/reveal traits other than sadism in someone (e.g., anger; dominance) but that make them more likely to act on their potential sadistic preferences and/or make them stronger.

(d) Glorification of suffering (“no pain no gain”, “suffering gives meaning to life”, etc.).

(e) Absence of factors making someone develop their empathy and emotional intelligence.

Children who enjoy torturing animals are likely to be victims of abuse themselves (Fiester and Gay 1991; McEwen et al. 2014; Spitzer et al. 1991). Also, Vizard et al (2007) suggest that children raised in what they consider “dysfunctional family environments” (defined as less specific than “abuse”, including “family breakdowns”) are more likely to display sadistic behaviors. This seems to be fairly strong evidence for at least something close to (a) and/or (c). It also doesn’t exclude (b) and (e) as potentially relevant here.

“Repeated exposure to violent video games predicted everyday sadism over time” in a study run by Greiteneyer and Sagioglou (2017), which is solid evidence for something in the vicinity of (a) and/or (c), and keeps the possibility of (d) pretty alive.

The strong correlation found between sadism and other Dark Tetrad traits (see e.g. Plouffe et al., 2018; Book et al. 2016; Tran et al. 2018; and references therein) as well as with other anti-social behaviors (see e.g. Pfattheicher et al. 2019; Rogers et al. 2018) is also some evidence for (c).

4. Directions for further research

• Are there important/interesting manifestations of sadism I omitted in §2?
• Can we come up with different – or more specified – theories for why different forms of sadism emerged if we approach the topic with different typologies than the one I use in §3.1 (e.g. Millon’s sadism subtypes)? Is that insightful at all?
• What can we learn from the literature on, e.g., sadistic personality disorder? Is there anything important we should know about how various forms of sadism work? For example,
  • is there any research not referenced in §3 that can help us further answer the questions this section asks?
  • how does sadism operate neurologically and how can this knowledge help us? (see e.g. Foulkes 2019, sections 5.3 & 5.4; and references therein).

Acknowledgment

I thank Manon Gouiran for her insightful comments on a draft for this post. I have also benefited from presenting and discussing the topic with the participants of the first S-Risk Introductory Fellowship, organized by the Center for Reducing Suffering. My work on the subject has been supported by a grant from Macroscopic Ventures in 2024.

All assumptions/claims/omissions remain my own.

Appendix A: Some correlations between sadism and some non-dark-tetrad traits

Misanthropy has been shown to correlate with some forms of sadism by, e.g., Asim et al. (2020, Figure 1). Similar results have been found for high risk tolerance by Nott and Walker (2021), as well as for spite by Lasko (2021),^[9] and assuming that these results somewhat generalize to things such as ideological fanaticism wouldn’t seem crazy. Sadism arguably also significantly correlates with lock-in-conducive preferences (see Althaus and Baumann 2020; Buhler 2023, section Malevolent actors are more likely to trigger a value lock-in; Finnveden et al. 2022; MacAskill 2022, Chapter 4), and a “motivation to undermine social structure” (see Paulhus et al. 2021).

Appendix B: Sadism in non-human animals?

For some behavior to count as sadism according to my definition, the perpetrator needs to understand their target is experiencing suffering and value that specifically (rather than value dominating them or just "playing around"). It seems hard to argue that non-human animals can meet that bar. A cat torturing birds by "playing with them" probably doesn't actually care about the fact that the bird is suffering.

However, if the evolutionary explanations are true, sadism in humans seems to be simply a slight value drift from these non-sadistic reasons to want something that implies suffering. And as I argued at the beginning of section 2, even in humans, many acts we think of as "sadistic-looking" are actually more of that kind, and maybe not actually motivated by sadism at all.

Also, given what I've said here, it wouldn't be surprising if sadism existed in some other mammals with high emotional intelligence? This would be very hard to prove, though. It is already not obvious with humans.

1. ^{^}

    Lucy Foulkes (2019, section 2.1 & 2.2) raises similar points, focusing on the difference between enjoying someone’s suffering and enjoying the associated domination.

2. ^{^}

    See respectively Paulhus et al. (2020) and O’Meara et al. (2011).

3. ^{^}

    See e.g. Paulhus et al. (2020); O’Meara et al. (2011).

^{^}

 Prominent well-documented and heart-shattering examples include the Murder of Junko Furuta and the many acts of kidnapping and torture of David Parker Ray. (I do not recommend learning more about these harrowing stories, honestly.)

^{^}

 “Even though deterrence has been proposed as a major reason for the evolution of retaliatory aggression, it is unclear whether people actually take revenge with deterrence in mind (see Osgood 2017). When explicitly asked to justify their revenge, people will cite deterrence motives (Darley & Pittman 2003) and will report feeling better about revenge when it has affected a positive change (Funk et al. 2014). However, there is also evidence that these self-reports are post hoc rationalizations rather than true motives. For example, when people are calculating the appropriate severity of retributive punishment, they are more attuned to whether the punishment matches the original transgression than whether it deters future harm (Carlsmith & Darley 2008). People playing economic games will also take revenge when they know they will not encounter their partner again, which would not make sense if revenge was solely intended to deter future harm (Fehr & Gachter 2000). Interestingly, vengeful people often feel less safe from future harm ¨ than do nonvengeful people (Akın & Akın 2016), indicating that people do not commonly take revenge because they think it will protect them. This evidence suggests that, of the many proximal predictors of revenge, deterrence may be among the least influential.” (Jackson et al. 2019)

^{^}

 This is (slightly) different from (pure) spite (see Macé et al. 2023) where the perpetrator terminally values frustrating someone’s preferences (rather than instrumentally because they intrinsically value their suffering). But sadism and spite are of course very related concepts. Sadism can be categorized as a specific version of spite and we’ve just seen that (pure) spite is barely different from spiteful sadism, a specific version of sadism.

^{^}

 More specifically, Paulhus and Dutton (2016) define everyday sadism as a set of “largely acceptable forms of subclinical sadism that are prevalent in modern culture”.

^{^}

 As Manon Gouiran rightfully pointed out while commenting on a draft, sadism does not need to have been evolutionarily selected in order to exist. However, there are some good reasons why some forms of sadism would have evolved, and it'd be very surprising if it was just a random variation, especially given that we can find some weak forms of sadism in almost all humans (rather than in a minority like fore, e.g., color blindness, which is plausibly due to random variations).

^{^}

 This isn’t surprising at all for spite. Spite is barely different from what I call “spiteful sadism” (a specific version of sadism) in §2.

Show all footnotes