1. Introduction
“The heat made people crazy. They woke from their damp bed sheets and went in search of a glass of water, surprised to find that when their vision cleared, they were holding instead the gun they kept hidden in the bookcase.” – Kristin Hannah, Summer Island: A novel (Reference Hannah2004)
Aggression and violence tear the fabric of society. They often pose a threat to feelings of safety and trust, undermine healthy relations among people, and bring about considerable suffering and unnecessary loss to people in many countries (Anderson Reference Anderson2001; Hsiang et al. Reference Hsiang, Burke and Miguel2013; Van de Vliert Reference Van de Vliert2009). One of the many factors that can make people more aggressive and violent is heat, as suggested by the opening quote from Kristin Hannah's book. One major scientific puzzle derives from the observation that the prevalence of aggression and violence differs within and between countries. As a general trend, aggression and violence increase as distance to the equator decreases (e.g., Walker et al. Reference Walker, Wilson, Chappell and Weatherburn1990). These differences are large and widespread. For example, data from the 2013 Global Study on Homicide (United Nations Office on Drugs and Crime [UNODC] 2013) reveal that, per 100,000 people, the rates for homicide are higher for Central America (26 per 100,000) and Middle Africa (18 per 100,000) than for Europe (5 per 100,000) and Northern America (5 per 100,000). There are, however, exceptions to this general “rule.” For example, although South Africa is quite distant from the equator, it has a very high violent crime rate (30 per 100,000). Violent crime differences also occur within continents. For example, differences in violent crime rates occur along the north-south axis in Europe, with homicide rates ranging from about 4 per 100,000 in Albania, Montenegro, and Turkey to less than 1 per 100,000 in Scandinavia. For within-continent comparisons there are exceptions as well, most notably Russia, with a homicide rate of at least 5 per 100,000.
Violent crime differences also occur within countries. The Federal Bureau of Investigation (FBI) has consistently reported that, in the United States, there is more violent crime in the South than in the North (FBI 2015). Similarly, the rate of Mafia-related homicides is much higher in southern than in northern Italy (UNODC 2013). Beginning at the global level and ending at the subnational level, whether across regions, subregions, or countries, two robust trends are observed with respect to aggression and violence: (1) there are significant differences between countries (and sometimes within countries), and (2) there tends to be more aggression and violence closer to the equator than further from the equator.
This bigger picture is supported in a recent meta-analysis on climate and conflict (see Burke et al. Reference Burke, Hsiang and Miguel2015) that revealed that climate is associated with violence in 46 of 56 (82%) published studies. Moreover, temperature is associated with violence in 20 of 24 studies (83%). Burke et al. also reported that effects were stronger for temperature than for rainfall differences and for intergroup than for interpersonal conflict. This meta-analysis provides a strong – and interdisciplinary – empirical foundation for the conclusion that “large variations in climate can have large impacts on the incidence of conflict and violence across a variety of contexts” (Burke et al. Reference Burke, Hsiang and Miguel2015, p. 610).
Although there are large differences in aggression and violence within and across countries, theoretical explanations for these differences are lacking. Most explanations focus on the influence of average temperature as a factor that triggers aggression and violence (General Aggression Model), or the notion that warm temperature allows for more social interaction situations (Routine Activity Theory) in which aggression and violence are likely to unfold. We propose a new model of CLimate, Aggression, and Self-control in Humans (CLASH) that helps us to understand differences within and between countries in aggression and violence in terms of differences in climate. Specifically, we propose that higher average temperature, and especially smaller seasonal variation in temperature, calls for individuals and groups to adopt a faster life strategy, a greater focus on the present (vs. future), and a lesser focus on self-control. The CLASH model further outlines that fast life strategy, short-term orientation, and lack of self-control are important determinants of aggression and violence.
Throughout this article, we use the terms aggression and violence to describe broad classes of behavior intended to harm others. Aggression is defined as any behavior that is intended to harm another person who is motivated to avoid that harm, and violence is defined as any behavior that is intended to cause extreme physical harm (e.g., injury, death) to another person who does not want to be harmed (cf. Anderson & Bushman Reference Anderson and Bushman2002). All violent acts are aggressive, but not all aggressive acts are violent – only acts intended to cause extreme physical harm are classified as violent. Also, our focus is not limited to acts of interpersonal aggression and violence. We also include acts of intergroup conflict, such as political violence, wars, and riots (see Burke et al. Reference Burke, Hsiang and Miguel2015).
As will be discussed, our conceptualization focuses on life strategies, time orientation, and self-control as constructs that are key to understanding aggression and violence. Each of these variables is shaped by climate (e.g., differences in average temperature, seasonal variation in temperature). Moreover, self-control in particular is assumed to be a powerful predictor of aggression and violence. Indeed, poor self-control is one of the “strongest known correlates of crime” (Pratt & Cullen Reference Pratt and Cullen2000, p. 952), especially violent crime (Gottfredson & Hirschi Reference Gottfredson and Hirschi1990; Henry et al. Reference Henry, Caspi, Mofitt and Silva1996). We therefore focus on those forms of aggression and violence that are due to low self-control. Specifically, we focus on “hot,” impulsive, angry behavior intended to harm another person who does not want to be harmed, called reactive aggression (also called hostile, affective, angry, impulsive, or retaliatory aggression [e.g., Buss Reference Buss1961; Dodge & Coie Reference Dodge and Coie1987]). Reactive aggression is liable to occur in situations where, for example, time to think is limited, cognitive load is high, immediate retaliation is feasible, and there is “a sense of urgency” to respond (e.g., in response to public humiliation, in direct confrontations). Reactive aggression can be a criminal act (e.g., assault, murder) or a noncriminal act (e.g., swearing at a rude driver, screaming at one's spouse).
Before we discuss our model in greater detail, we should explicate three foci of the present theoretical analysis. First, we acknowledge that comparisons within countries are less complex than comparisons between countries. In general, there are far fewer differences within countries than between countries. Differences between countries (e.g., historical, economic, political variables) are exceptionally difficult to disentangle from climate differences (cf. Burke et al. Reference Burke, Hsiang and Miguel2015). Thus, our analysis focuses more on within-country than on between-country comparisons of climate differences. We acknowledge that between-country comparisons are important with respect to the scientific principle of efficiency (i.e., explaining a lot of variance using a relatively parsimonious model) and the societal urgencies the world faces (e.g., global change, migration issues, and international cooperation [cf. Van Lange Reference Van Lange2013]).
Second, we focus on the Northern Hemisphere for methodological and practical reasons. A large majority of the world's population live in the Northern Hemisphere. Moreover, past research has focused on countries in the Northern Hemisphere. It is this past research that is in strong need of a new model able to account for pronounced differences between southern and northern environments (in the Northern Hemisphere) on several important dimensions: time orientation, self-control, aggression, and violence. Theoretically, the distance from the equator should work the same way in the Southern Hemisphere as in the Northern Hemisphere. Indeed, in the final analysis, we believe that it is desirable, from a scientific and societal perspective, to extend the model to both hemispheres (cf. Henrich et al. Reference Henrich, Heine and Norenzayan2010).
Third, as the name conveys, CLASH focuses on humans, rather than other animals. We acknowledge that animals also adapt and respond to climatic differences (see Burghardt Reference Burghardt2013). For example, climate differences are associated with hibernation and storage of food for some species (e.g., bears, skunks, and chipmunks), migration to other regions for some species (e.g., fish, birds, and butterflies), and movement to specific locations such as underground burrows or holes in trees in the same region for some species (e.g., mice, snakes, and frogs). Climate is also linked to seasonal “planning” of reproduction for many species that mate in the Spring (Wikelski et al. Reference Wikelski, Hau and Wingfield2000). Although these patterns of adaptation can be viewed in terms of life history strategies, time orientation, and self-control, we believe it is premature to link these patterns to aggression and violence in other animals, for two reasons. First, we do not know of any empirical literature on the link between climate differences and aggression among the same animal species. It is more likely that throughout evolutionary history, animals have either adapted to the local climatological circumstances or migrated to more fitting circumstances. These topics are beyond the scope of this article. Second, as illustrated above, many species have their own unique way of adapting to annual differences in seasonal climate. This is not to imply that we regard an examination of comparative research as unimportant. Indeed, we hope that the specific tests of CLASH may be extended to humans and other animals in the future.
The remainder of the article is organized as follows. Section 2 presents the existing evidence and theories linking climate to aggression and violent crime. Before we describe our CLASH model, it is necessary to provide a brief overview of theories and research relevant to differences in aggression and violence, both within and between countries. This is especially important in outlining what CLASH contributes to our current understanding of violence within and across countries. Section 3 discusses theory and research relevant to two propositions that provide the foundation for CLASH. Given that climatological approaches are not common in the social and behavioral sciences, Section 4 includes a broad discussion about the ubiquity of climate for understanding human behavior in groups and societies. Because CLASH offers a novel and general framework, Section 5 considers caveats and future directions of CLASH. Section 6 concludes by outlining theoretical issues and broader scientific and societal implications of CLASH. The final section includes some concluding comments.
2. Contemporary explanations of cultural differences in violence
Inspired by the observation that countries closer to the equator are generally more violent, several theorists and researchers have attempted to determine why there is so much variation in aggression and violence around the world. One belief shared by experts and laypeople alike is that hot temperatures increase violence. The belief that higher temperatures increase violence has spurred researchers to examine the role of average heat (climate) and incidental heat (weather) in violence rates since the late 1800s (e.g., Dexter Reference Dexter1899; Lombroso 1899/Reference Lombroso1911; for comprehensive reviews, see Brearley Reference Brearley1932; Cohen Reference Cohen1941; Falk Reference Falk1952).
Considerable research has indicated that as temperature increases, violent crime (e.g., murder, rape, assault, violent riots) also increases (Anderson Reference Anderson1987; Reference Anderson1989; Carlsmith & Anderson Reference Carlsmith and Anderson1979; deFronzo Reference deFronzo1984; Michael & Zumpe Reference Michael and Zumpe1986), but there is no corresponding increase in nonviolent crimes. Also, as noted earlier, a variety of studies conducted in the United States have found that Southern states with warmer climates typically have higher violent crime rates than Northern states with cooler climates (e.g., Anderson & Anderson Reference Anderson and Anderson1996; Lombroso 1988/Reference Lombroso1911; UNODC 2013). Similarly, time period studies on temperature variability have revealed higher violent crime rates in hotter years, seasons, months, and days (e.g., Anderson et al. Reference Anderson, Bushman and Groom1997; Leffingwell Reference Leffingwell1892). In addition, field and archival studies have found a positive correlation between heat and aggression in a variety of forms (e.g., horn honking, number of major league baseball batters hit by pitched balls, prison inmate violence [Haertzen et al. Reference Haertzen, Buxton, Covi and Richards1993; Kenrick & MacFarlane Reference Kenrick and MacFarlane1984; Reifman et al. Reference Reifman, Larrick and Fein1991]). Overall, correlational studies, field experiments, and archival studies of violent crimes provide evidence of the “heat effect” – higher temperatures are associated with higher levels of aggression and violence.
Given that various empirical studies have reported that as temperature increases, so do aggression and violence, the obvious question is: What is it about high temperatures that makes people generally more aggressive and violent? The two most popular theories offered to account for this positive relation between temperature and violence are the General Aggression Model (Anderson & Bushman Reference Anderson and Bushman2002) from psychology and the Routine Activity Theory (Cohen & Felson Reference Cohen and Felson1979; Rotton & Cohn Reference Rotton and Cohn2001) from law and criminology.
2.1. General Aggression Model
In the General Aggression Model (e.g., Anderson & Bushman Reference Anderson and Bushman2002), two types of input variables can influence whether a person acts aggressively: personal variables (e.g., genetic predispositions, trait aggression, gender, attitudes about violence) and situational variables (e.g., alcohol, violent media, provocation, hot temperatures). The relevant situational variable here is temperature. According to the model, there are three possible routes to aggression and violence: angry feelings, aggressive thoughts, and physiological arousal. Together these three routes constitute an individual's present internal state, which encourages or discourages aggression and violence. However, these routes are not mutually exclusive or even independent. For example, someone who has aggressive ideas might also feel angry and have elevated blood pressure. High temperatures appear to operate through all three routes. For example, high temperatures make people angry, increase aggressive thoughts, and increase physiological arousal (e.g., heart rate, blood circulation, perspiration). This unexplained arousal by the heat can be mislabeled as “anger,” especially in situations involving provocation and thus lead to reactive aggression (Zillmann Reference Zillmann1979). This might help explain why a minor provoking social event, such as an accidental bump in a hot crowded bar, can lead to the trading of insults, punches, and possibly even bullets (Anderson Reference Anderson2001).
2.2. Routine Activity Theory
The Routine Activity Theory (Cohen & Felson Reference Cohen and Felson1979; Rotton & Cohn Reference Rotton and Cohn2001) conceptualizes the effect of weather on violent crime rates in terms of the amount of social contact. As one scholar wrote, “the greater frequency of crimes against the person in summer months is probably due to the greater frequency of contact among human beings in those months rather than the effects of temperature on the propensity to criminality” (Sutherland & Cressey Reference Sutherland and Cressey1978, p. 119). The rationale underlying the Routine Activity Theory is relatively straightforward – during warmer weather, individuals are more likely to leave the safety of their homes, schools, and jobs and spend more time outside in public spaces, where interactions with others can become “heated” and aggressive (Cohn Reference Cohn1990). Consistent with these predictions, violent crime rate data from Minneapolis, Minnesota, and Dallas, Texas, indicate that the relation between hot temperatures and violent crime is stronger when individuals spend generally more time outdoors in the evening rather than afternoon hours and on weekends rather than weekdays (e.g., Cohn & Rotton Reference Cohn and Rotton1997; Rotton & Cohn Reference Rotton and Cohn2001).
2.3. Hot or not: Past theory and research
Despite the wealth of empirical studies on the heat effect, three limitations are worth mentioning. First, although the General Aggression Model (Anderson & Bushman Reference Anderson and Bushman2002) proposes that heat-induced anger, aggressive thoughts, and physiological arousal can lead to more aggression, it is unlikely that these factors alone would lead to extremely violent behaviors such as homicide. Indeed, there is evidence that the effect sizes of heat-induced hostility are relatively modest, both inside and outside of the lab (e.g., Ferguson & Dyck Reference Ferguson and Dyck2012). Moreover, laboratory experiments have yielded mixed results. Some experiments indicate that extremely hot temperatures inhibit aggression, presumably because people want to escape the heat rather than fight (e.g., Baron Reference Baron1972; Baron & Bell Reference Baron and Bell1975; Reference Baron and Bell1976). Also, some evidence suggests that aggression and violence occur less frequently in hot climates than in warm climates (e.g., Van de Vliert et al. Reference Van de Vliert, Schwartz, Huismans, Hofstede and Daan1999). Taken together, past research suggests that a greater scientific understanding of the mechanisms underlying the relationship between heat and aggression is needed (see Anderson & Anderson Reference Anderson, Anderson, Geen and Donnerstein1998).
Second, although the Routine Activity Theory proposes that the link between temperature and crime is due to individuals' congregating in public spaces with increased social interaction, this perspective has not always received empirical support (Rotton & Cohn Reference Rotton and Cohn2000). For example, although there is a greater likelihood of violent behaviors among young people in a bar room setting, violent behaviors are unlikely to occur in this setting among groups of mixed ages and sexes (Felson Reference Felson1998). This observation is consistent with what is known in the criminology literature as the “nighttime economy,” which consists primarily of bars, pubs, and nightclubs, settings in which alcohol-related violence can occur (Teece & Williams Reference Teece and Williams2000). For example, the correlation between hot temperature and violent crime is generally stronger during weekend evenings and nighttime hours when the temperatures are cooler and around pubs and nightclubs (e.g., Allen et al. Reference Allen, Nicholas, Salisbury, Wood, Flood-Page and Taylor2003; Bushman et al. Reference Bushman, Wang and Anderson2005; Tierney & Hobbs Reference Tierney and Hobbs2003). Another study found that robbery rates tend to increase in the evening during the fall when the sun sets earlier and tend to decrease in the spring when the sun sets later (Doleac & Sanders Reference Doleac and Sanders2013). Moreover, congregation of a handful of individuals in public places (e.g., during festivals) can lead to identification and social cohesion (Whitehouse & Lanman Reference Whitehouse and Lanman2014). All else being equal, social contact appears to be a necessary, but not a sufficient condition for the occurrence of violent crime. In addition, much violence occurs inside of the home among family members and close friends, rather than outside of the home (e.g., DeWall et al. Reference DeWall, Lynch, Renzetti and Bushman2016; Krahé Reference Krahé and Bushman2017).
Third, various studies that have examined climate differences and violence around the world have included countries with high average temperatures and small seasonal variation (e.g., India, Indonesia, Kenya, and sub-Saharan Africa [see Burke et al. Reference Burke, Hsiang and Miguel2015; Simister & Van de Vliert Reference Simister and Van de Vliert2005; Van de Vliert Reference Van de Vliert2009]). However, many studies on the association between temperature and aggression or violence, especially field studies and laboratory experiments, have been conducted in the United States. But even within the United States, the General Aggression Model and the Routine Activity Theory cannot explain some other violence-relevant attitudes that are quite different for most states in the South versus those in the North. For example, in most Southern states, there is greater approval and support for corporal punishment, gun ownership, and capital punishment than in most Northern states (Shackelford Reference Shackelford2005). Several scholars have argued that pro-violence attitudes in the South are characterized by “machismo,” masculine aggression (Simister & Van de Vliert Reference Simister and Van de Vliert2005), or a “Southern culture of honor,” an ideology justifying the use of violence for self-defense and defense of one's “honor” or reputation for being strong, tough, brave, and manly (e.g., Cohen Reference Cohen1996; Reference Cohen1998; Cohen & Nisbett Reference Cohen and Nisbett1994; Reference Cohen and Nisbett1996; Cohen et al. Reference Cohen, Nisbett, Bowdle and Schwartz1996; Reference Cohen, Vandello, Puente and Rantilla1999; Nisbett Reference Nisbett1993; Nisbett & Cohen Reference Nisbett and Cohen1996).
Some scholars have proposed that the Southern culture of honor in the United States developed in response to the herding economy of the frontier region of the South (Nisbett & Cohen Reference Nisbett and Cohen1996). Because herding (more than farming) places an individual at risk for losing everything from theft, and because the South was a frontier region where the state had little power to prevent or punish theft of property, individuals created and enforced their own system of law and order defined by “the rule of retaliation.” However, it is not clear why the Southern culture of honor still exists today, especially because the modern South is not based on a herding economy and is not lawless (Shackelford Reference Shackelford2005). It is possible that the psychological mechanisms underlying the behavioral manifestations of the Southern culture of honor were selected as a solution to some other adaptive problem characteristic of the South (vs. North).
To summarize, although contemporary explanations of cultural differences in violence provide compelling cultural accounts of violence, they have conceptual and methodological limitations. The explanations focus more on behavioral patterns than on underlying mechanisms. The culture of honor hypothesis focuses on historical determinants in particular regions of the United States. Perhaps most importantly, the explanations do not account for the climate differences that may underlie the exceptionally large and widespread differences in violent crime within and between various countries around the world.
3. CLASH: A model of CLimate, Aggression, and Self-control in Humans
Through our CLASH model (see Fig. 1), we seek to explain differences within and between countries in terms of temperature, especially seasonal variation in temperature. Using an extension of Life History Theory and the broader literature on time orientation and self-control, we advance two propositions suggesting that temperature-related aggression and violence can be understood in terms of time orientation and self-control. Although theoretical in nature, the propositions are rooted in research conducted in various disciplines of the social and behavioral sciences, with an emphasis on social and evolutionary psychology.
Figure 1. A model of CLimate, Aggression, and Self-control in Humans (CLASH)
Our proposals are organized around calls for the development of more interdisciplinary theories. Three broad categories of factors that influence aggression and violence levels in countries and regions are: (1) climatological, (2) evolutionary, and (3) psychological. Our goal is not to exhaustively catalogue the many factors that influence aggressive and violent behaviors. Rather, we seek to advance the theoretical understanding of the pronounced differences in aggression and violence within and between countries around the world. Unlike other explanations that focus primarily on average differences in climate (hot vs. cold climates), we focus on average temperature and especially the broad influence of seasonal variation in climate (small or big annual differences within a location) on life strategy, time orientation, self-control, aggression, and violence. Thus, our CLASH model provides a novel perspective from which to observe why countries and regions closer to the equator tend to have higher levels of aggression and violence than do countries farther away from the equator.
The key climatological variables that influence aggression and violence are average temperature and seasonal variation in temperature. Of course, climate also entails such variables as rainfall, wind, water availability, and climate indices (e.g., El Niño Southern Oscillation Index). We focus on temperature for three reasons. First, the extant body of research has examined primarily temperature as the key climatological variable (e.g., Van de Vliert Reference Van de Vliert2013a). Second, a recent meta-analysis has revealed that the association between temperature and conflict is at least four times as strong as the association between rainfall and conflict (Burke et al. Reference Burke, Hsiang and Miguel2015). Third, for most countries, there is greater predictable seasonal variation in temperature than in other climatological variables such as rainfall. Thus, although we share the view that climate differences in terms of averages and seasonal variability differ in several interesting respects, we focus on temperature rather than precipitation as the key variable.
In the next section, we discuss in detail two propositions that provide the foundation for our CLASH model. One broad assumption of CLASH is that adaptation to various climates is reflected in slow and fast life strategies, in differences in time orientation and self-control, and in differences in aggression and violence levels. Proposition 1 states that lower temperatures and especially greater seasonal variation in temperature call for individuals and societies to adopt a slower life strategy, a greater future orientation, and greater self-control. Proposition 2 states that lower temperatures and especially greater seasonal variation in temperature help individuals and societies evolve to be less aggressive and less violent in situations requiring future orientation and self-control. As with all scientific propositions, these propositions are subject to revision, refinement, and progress. Our primary goal in formulating CLASH is to propose a new theoretical model, and the propositions should help researchers develop and test specific hypotheses relevant to CLASH.
3.1. Proposition 1
One key lesson of evolutionary theory is that resources for survival and reproduction are not infinite. Hence, a basic challenge to all organisms is the successful allocation of resources needed for survival and reproduction. Natural selection favors resource allocation strategies that, in response to environmental conditions, enhance an organism's inclusive fitness over the life span (Ellis et al. Reference Ellis, Figueredo, Brumbach and Schlomer2009).
A prominent theory that focuses on how different resource allocation strategies arise from different exposures to environmental conditions is Life History Theory (Hill Reference Hill1993; Kaplan & Gangestad Reference Kaplan, Gangestad and Buss2005; MacArthur & Wilson Reference MacArthur and Wilson1967; Pianka Reference Pianka1970). This theory concerns the allocation of finite resources across different fitness-relevant activities. According to some theorists, two features of an environment are essential for psychological development and adaptation: harshness and unpredictability (Ellis et al. Reference Ellis, Figueredo, Brumbach and Schlomer2009; Griskevicius et al. Reference Griskevicius, Tybur, Delton and Robertson2011). Harshness refers to the rates of mortality and morbidity caused by largely uncontrollable factors (e.g., high rates of infectious disease [Frankenhuis et al. Reference Frankenhuis, Panchanathan and Nettle2016]). Unpredictability refers to the uncertainty of future outcomes. The environmental threats of harshness and unpredictability, in combination with the resources available for coping with environmental threats, largely determine how stressful an environment is. These are features that are often reflected in higher morbidity and mortality (Adler et al. Reference Adler, Boyce, Chesney, Cohen, Folkman, Kahn and Syme1993; Chen et al. Reference Chen, Matthews and Boyce2002).
Life History Theory proposes that people adapt to (un)harshness and (un)predictability by adopting either a fast (slow) life history strategy. Relative to slow life strategies, fast life history strategies are associated with reproducing at an earlier age, having more uncommitted and less stable sexual relationships, having more children, and investing less time, effort, and resources in each child. Also, relative to slow strategies, fast life strategies tend to be associated with short-term planning, greater risk taking, a focus on immediate gratification for short-term benefits, and more aggression (e.g., Ellis et al. Reference Ellis, Figueredo, Brumbach and Schlomer2009; Frankenhuis et al. Reference Frankenhuis, Panchanathan and Nettle2016; Griskevicius et al. Reference Griskevicius, Tybur, Delton and Robertson2011; Nettle Reference Nettle2010; Simpson et al Reference Simpson, Griskevicius, Kuo, Sung and Colling2012). Thus, Life History Theory posits that people adapt to harsh and unpredictable environments by adopting faster life strategies. Because the future is unpredictable and people tend to die sooner in such environments, it is adaptive for people to enact fast life strategies because delayed payoffs may never be realized. In contrast, in environments that are unharsh and predictable, people adopt slower life strategies. Because the future is more predictable and people tend to live longer in such environments, it is adaptive for people to enact slower life strategies because delayed payoffs are likely to be realized (Ellis et al. Reference Ellis, Figueredo, Brumbach and Schlomer2009; Griskevicius et al. Reference Griskevicius, Tybur, Delton and Robertson2011; Simpson et al. Reference Simpson, Griskevicius, Kuo, Sung and Colling2012).
An abundance of research has supported this view. As noted earlier, threats of harshness and unpredictability are often reflected in higher morbidity and mortality. Lower socioeconomic status (SES) is related to nearly all forms of morbidity and mortality (Adler et al. Reference Adler, Boyce, Chesney, Cohen, Folkman, Kahn and Syme1993; Chen et al. Reference Chen, Matthews and Boyce2002; Miller et al. Reference Miller, Chen and Parker2011). From a life history perspective, one might expect that low-SES individuals should enact faster life strategies than high-SES individuals because they are more likely to suffer premature disability and death (Adler et al. Reference Adler, Boyce, Chesney, Cohen, Folkman, Kahn and Syme1993; Chen et al. Reference Chen, Matthews and Boyce2002; Miller et al. Reference Miller, Chen and Parker2011). Indeed, lower SES is associated with a number of fast life strategies, such as earlier sexual activity (e.g., Ellis et al. Reference Ellis, Bates, Dodge, Fergusson, Horwood, Pettit and Woodward2003; Kotchick et al. Reference Kotchick, Shaffer, Forehand and Miller2001), higher rates of adolescent pregnancy and childbearing (e.g., Ellis et al. Reference Ellis, Bates, Dodge, Fergusson, Horwood, Pettit and Woodward2003; Miller et al. Reference Miller, Benson and Galbraith2001), greater number of offspring (Vinning Reference Vinning1986), and lower levels of parental investment per child (e.g., Belsky et al. Reference Belsky, Steinberg and Draper1991; Ellis et al. Reference Ellis, McFayden-Ketchum, Dodge, Pettit and Bates1999).
Similar observations can also be made for other factors in environmental harshness and unpredictability. For example, past research has indicated that there is a greater likelihood for individuals growing up in harsh and unpredictable family environments (e.g., homes with a lot of fighting between family members) to experience faster sexual maturation, earlier age of reproduction, and higher reproductive rates (e.g., Chisholm Reference Chisholm1999; Kim et al. Reference Kim, Smith and Palermiti1997). Moreover, neighborhood deterioration and danger (e.g., assaults, muggings, burglaries, thefts, presence of gangs and drug addicts) are associated with earlier sexual activity and higher rates of risky sexual behaviors (e.g., Lauritsen Reference Lauritsen1994; Upchurch et al. Reference Upchurch, Aneshensel, Sucoff and Levy-Storms1999). Furthermore, as resources become increasingly scarce, females increasingly prefer mates who have access to resources, and parents increasingly invest in their offspring's reproductive value (e.g., Bugenthal & Beaulieu Reference Bugenthal and Beaulieu2004; Durante et al. Reference Durante, Griskevicius, Redden and White2015; Kruger et al. Reference Kruger, Reischl and Zimmerman2008).
Our CLASH model extends Life History Theory. In particular, Life History Theory emphasizes unpredictability and harshness as sources of environmental stress, whereas CLASH emphasizes predictability as a source of control over environmental stress (see also Ellis et al. Reference Ellis, Figueredo, Brumbach and Schlomer2009). By control we mean the actions that can be taken to adapt optimally to predictable change, especially in preparation for predictable harsh circumstances. Although control is always low in unpredictable situations, it can be high in predictable situations. CLASH proposes that the combination of predictability and control shape a slow life strategy, a future time orientation (e.g., an orientation relevant to planning purposes), and a focus on self-control (to control short temptations and pursue long-term goals).
CLASH proposes that greater distance from the equator is associated with a slower life strategy, a stronger future orientation, and a greater focus on self-control. The key explanatory variables are average temperature and seasonal variation in temperature (see Fig. 1). In regions closer to the equator, the climate is warmer and less variable per season, and so individuals have less of a need to plan ahead to ensure survival and reproduction. That is, there is little need to focus on the future, develop a longer time perspective (Kruger et al. Reference Kruger, Reischl and Zimmerman2008), or exercise self-control (Baumeister et al. Reference Baumeister, Park and Ainsworth2013). Moreover, societies closer to the equator are also relatively harsh and unpredictable. Hot temperatures can be an important source of stress, not only in terms of everyday life, but also as a threat to harvests in agriculture. Another source of harshness and unpredictability is pathogen stress. Indeed, the prevalence of parasitic and infectious diseases, such as malaria and the Zika virus, is considerably higher in countries closer to equator (e.g., Guernier et al. Reference Guernier, Hochberg and Guégan2004), which poses a threat to survival and may activate human affect, cognition, and behavior, such as direct vigilance, stress, and escape (e.g., Fincher & Thornhill Reference Fincher and Thornhill2012; Fincher et al. Reference Fincher, Thornhill, Murray and Schaller2008). Also, there is some evidence that the risk of natural disasters tends to increase as distance to the equator decreases (National Oceanic and Atmospheric Administration 2016).
In societies more distant from the equator, people face both lower temperatures and greater seasonal variation in temperature. Both characteristics, but especially seasonal variation in temperature, should give rise to a slower life strategy, a stronger future orientation, and a stronger focus on self-control. Although there is some harshness in these societies, there is also predictability – events are largely controllable in terms of planning and “coping.” In particular, individuals in these societies realize that they need to plan and prepare for the next season. For example, food supply is less plentiful and less varied during winter, posing a serious threat to health. Yet the quality and quantity of food supply can be promoted by adopting a future orientation (e.g., planning) and by exercising self-control (resisting the temptation to consume the harvest directly, a commitment to work hard to optimize the harvest for later [Ainslie Reference Ainslie2013; Baumeister et al. Reference Baumeister, Park and Ainsworth2013]). Indeed, an analysis of work-related values in 40 countries revealed that countries located farther from the equator tend to place greater value on future-oriented rewards such as perseverance and thrift (Hofstede Reference Hofstede2001). In the next sections, we discuss empirical evidence relevant to CLASH.
3.1.1. Fast versus slow life strategy
According to CLASH, people in regions with lower temperatures and greater seasonal variation in temperature tend to adopt a slower life strategy. Distance from the equator is a good approximation for lower temperatures and greater seasonal variation in temperature. Consistent with our prediction, life expectancy is lower for countries closer to the equator than for countries farther from the equator. For example, in several African countries, Haiti, and Pakistan, life expectancy is lower than 65 years, whereas in many European countries and North America, life expectancy is higher than 80 years (World Health Organization 2013). Of course, there are some exceptions, most notably high-latitude countries near Russia (with life expectancies often lower than 70) and low-latitude countries such as Ecuador, Thailand, and Indonesia (with life expectancies of 70 or higher).
As noted earlier, one of the strongest and most objective indicators of slow versus fast life strategy is the mother's age at the birth of her first child. According to the World Factbook (Central Intelligence Agency 2014), the mother's age at first birth is less than 20 (on average) in countries closer to the equator (e.g., Gaza Strip, Liberia, Bangladesh, various middle African countries such as Kenya, Mali, Tanzania, Uganda). In contrast, mother's age at first birth is greater than 28 (on average) in countries further from the equator (e.g., Japan, Canada, and nearly all European countries). There are some exceptions to this general rule, such as Hong Kong and Singapore. Within the United States, a similar albeit less pronounced trend is observed (National Vital Statistics Reports). The five states with the lowest maternal age at first birth are located in the South: Mississippi (22.5), Arkansas (22.7), New Mexico and Louisiana (23.0), and Oklahoma (23.1). In contrast, the five states with the highest age at first birth are in the North: Massachusetts (27.8), Connecticut (age 27.2), New Jersey (27.1), New Hampshire (26.7), and New York (26.4).
Research also supports our hypothesis that a slow life history strategy is characterized by behaviors that reflect long-term planning, such as more restrictive reproductive behavior with greater parental investment in offspring. Under predictable environmental conditions, slower life history strategies would be better for enhancing an individual's inclusive fitness. Even when some harsh conditions become predictable, some control can be exerted by anticipating, preparing, and planning activities relevant to these conditions (e.g., Griskevicius et al. Reference Griskevicius, Tybur, Delton and Robertson2011). In these kinds of predictable and controllable environments, individuals contribute to their own embodied capital (e.g., growth and maintenance of their body and mind, accumulation of knowledge and skills [Mittal & Griskevicius Reference Mittal and Griskevicius2014]). Thus, there is growing evidence that predictable environments tend to promote a slower life strategy, in terms of lower mortality and morbidity, delayed reproduction, and higher contributions to one's own embodied capital.
3.1.2. Time orientation and self-control
One key assumption in CLASH is that the harshness and predictability of the environment influence time-orientation and self-control.
In this section, we review the empirical evidence relevant to similarities and differences within and between countries in terms of both time orientation and self-control. Before doing so, we outline the differences between these two concepts and then provide a brief general review of time orientation.
Time orientation is strongly connected to concepts such as “time perspective” and “temporal discounting.” It is also closely linked to self-control and related concepts such as delay of gratification and impulsivity. An orientation to the present is linked to lower levels of self-control than an orientation to the future (e.g., Baumeister et al. Reference Baumeister, Heatherton and Tice1994). However, it is important to distinguish between the broad concepts of time orientation and self-control. Self-control is generally conceptualized as the ability to resist and manage “temptations” and “impulses” (see Baumeister & Tierney Reference Baumeister and Tierney2011; Joireman et al. Reference Joireman, Balliet, Sprott, Spangenberg and Schultz2008), whereas time orientation is generally conceptualized as an orientation to the present versus the future (cf. Boniwell & Zimbardo Reference Boniwell, Zimbardo, Linley and Joseph2004; Joireman et al. Reference Joireman, Anderson and Strathman2003).
Time has objective or at least consensual features, such as “geography” and “clock time” (Boniwell & Zimbardo Reference Boniwell, Zimbardo, Linley and Joseph2004; Snyder & Lopez Reference Snyder and Lopez2009). Yet people experience time differently across countries around the world. For example, comparison of the United States with Brazil with respect to time reveals large differences (Levine Reference Levine2006). In the United States, the conception of time emphasizes the urgency of using time efficiently, making every minute count (Levine et al. Reference Levine, West and Reis1980). In contrast, in Brazil public clocks and personal timepieces often are intentionally set at different times (with differences up to 20 min), students often come late to class, and individuals often come late to formal appointments. Some of these differences may also be reflected in language. Countries farther from the equator emphasize the “extrinsic” value of time (e.g., “time is money”), whereas countries closer to the equator emphasize the “intrinsic” value of time. For example, in Mexico the phrase “give time to time” (darle tiempo al tiempo) is common; in Africa, the phrase “even the time takes its time” is common; and in Trinidad, the phrase “any time is Trinidad time” is common (Levine Reference Levine2006). Other scholars have distinguished between clock-time cultures and event-time cultures. Clock-time cultures are more future oriented than are event-time cultures. For example, the United States and Northern European countries are clock-time cultures that rely heavily on schedules and punctuality, whereas most countries in Latin America are event-time cultures that go with the natural flow of social events (Brislin & Kim Reference Brislin and Kim2003; Levine Reference Levine2006).
Some research has focused on “pace of life,” defined in terms of rapidity or density of experiences, perceptions, and activities (Werner et al. Reference Werner, Altman, Oxley, Altman and Werner1985, p. 14). A slower pace of life corresponds to a present orientation, whereas a faster pace of life corresponds to a future orientation. An analysis of individuals from 31 countries found that individuals from colder countries located further from the equator had a faster pace of life than did individuals from warmer countries located nearer the equator (Levine & Norenzayan Reference Levine and Norenzayan1999). Pace of life was measured using three behaviors: (1) the average walking speed of individuals, (2) the average time needed for a routine transaction in a post office, and (3) the average accuracy of public clocks. Another study comparing Fresno, California, with Niteroi, Brazil, found that public clocks and personal time pieces were less accurate in Brazil and that Brazilians were more likely to be late for appointments, were more flexible in their definitions of early and late, were less likely to attribute being late to internal factors, were less likely to express regret over being late, and were less likely than Americans to rate punctuality as an important characteristic in a businessperson or friend (Levine et al. Reference Levine, West and Reis1980). Niteroi, Brazil, is located much nearer the equator than Fresno, California.
Unfortunately, large cross-national studies on self-control are relatively sparse. Most studies on self-control are conducted in the United States, and if they are cross-national they often include countries from similar global regions. Also, some studies use domain-specific assessments of self-control (e.g., dieting) or antisocial behaviors that are not aggressive (e.g., truancy). One exception is a recent study that examined a self-report measure of self-control among children (Botchkavar et al. Reference Botchkavar, Marshell, Rocque and Posick2015). This study found higher levels of self-control in Northern European countries (e.g., Scandinavian countries, Iceland) than in Southern European countries or the United States. This finding, along with findings from cross-national studies on time orientation, provides some initial evidence for greater levels of self-control in countries farther from the equator.
3.1.3. Conclusions
Taken together, the empirical evidence supports the proposition that individuals and cultures are more likely to adopt a slower life strategy and to become more future oriented and less present oriented, as average temperatures decrease and seasonal variation in temperature increases. We should acknowledge that most studies involve comparisons among only a few countries, although a few studies have compared more than 20 countries. Moreover, various third variables may account for these differences. For example, there might be a positive association between a country's wealth or prosperity and future orientation (Milfont & Gapski Reference Milfont and Gapski2010). It is also noteworthy that the evidence typically yields support across a variety of indicators of time orientation and that very few studies yield conflicting evidence. Unfortunately, the “ideal” study remains to be conducted. Such a study would correlate distance from the equator and average and seasonal variation in temperature with time orientation and self-control. For a comprehensive test of Proposition 1, we recommend the use of self-report measures of both time orientation and self-control, but also instruments or assessments that do not rely on self-reports, such as unobtrusive behavioral measures. Thus, although conclusive evidence has not yet been obtained, the available evidence provides a relatively coherent picture that certainly is in line with Proposition 1 of CLASH (see Fig. 1).
3.2. Proposition 2
CLASH proposes that average temperature and seasonal variation in temperature have shaped the evolution and development of different adaptations in terms of life strategy, time orientation, and self-control. In this section, we discuss research on the link between temperature and seasonal variation in temperature and aggression and violence, along with the mediating roles of life strategy, time orientation, and self-control (see Fig. 1). That is, we advance the proposition that in regions with lower temperatures and greater variation in temperature, aggressive and violent behaviors are less likely because individuals have adopted a slower life strategy, a longer time orientation, and a higher level of self-control to adapt to their environment.
There is evidence that time orientation is linked to aggression and violence. Earlier research revealed that “delinquents” are more likely to think about the short-term than the long-term consequences of their actions (Gottfredson & Hirschi Reference Gottfredson and Hirschi1990; Pratt & Cullen Reference Pratt and Cullen2000). Other studies have found that experimental manipulations of “future self” reduce cheating in testing situations (Van Gelder et al. Reference Van Gelder, Hershfield and Nordgren2013; Reference Van Gelder, Luciano, Weulen Krananbarg and Hershfield2015). Also, several studies have investigated the role of time orientation in human cooperation, selfish behavior, and aggressive impulses. For example, people who are more prone to adopt a future orientation conserve natural resources (Kortenkamp & Moore Reference Kortenkamp and Moore2006), support structural solutions to social dilemmas (Van Lange et al. Reference Van Lange, Joireman, Parks and Van Dijk2013), and resist the urge to respond aggressively when insulted (Joireman et al. Reference Joireman, Anderson and Strathman2003).
The anticipation of future interaction is a powerful determinant of unselfish and cooperative behavior in social dilemmas (Van Lange et al. Reference Van Lange, Klapwijk and Van Munster2011). Similarly, adopting a long-time orientation in relationships inhibits selfish and retaliatory responses in close relationships (Rusbult & Van Lange Reference Rusbult and Van Lange2003). A future orientation is negatively correlated with trait aggressiveness (Joireman et al. Reference Joireman, Anderson and Strathman2003; Zimbardo & Boyd Reference Zimbardo and Boyd1999), hypothetical aggression in scenarios (Joireman et al. Reference Joireman, Anderson and Strathman2003), aggressive driving (Moore & Dahlen Reference Moore and Dahlen2008; Zimbardo et al. Reference Zimbardo, Keough and Boyd1997), and actual aggressive behavior, that is, willingness to administer electric shocks to another person in a laboratory experiment (Bushman et al. Reference Bushman, Giancola, Parrott and Roth2012). Thus, we conclude that a future orientation reduces selfish and aggressive behavior.
There is considerable research on the association between self-control and aggression and violence. In fact, one of the best predictors of violent criminal behavior is low self-control (see Gottfredson & Hirschi Reference Gottfredson and Hirschi1990). Indeed most murders committed in the United States are due to unchecked anger (FBI 2015). When angry feelings and violent urges become activated, self-control is what keeps them in check. Aggression often starts when self-control stops. Interestingly, experimental research has shown that self-control exercises can decrease aggression. In one experiment, for example, participants who had previously completed a measure of trait aggressiveness were randomly assigned to complete self-control exercises using their nondominant hand for everyday tasks (self-control training condition) or to answer math problems (control condition) for 2 weeks (Denson et al. Reference Denson, Capper, Oaten, Friese and Schofield2011). After 2 weeks, participants were provoked by a confederate in the laboratory and were given the opportunity to retaliate by administering aversive noise blasts to the confederate thorough headphones. The results indicated that the self-control exercises decreased aggression, especially for individuals high in trait aggressiveness. Another experiment found that partners who practiced self-control were less aggressive toward their loved one than were partners who did not practice self-control (Finkel et al. Reference Finkel, DeWall, Slotter, Oaten and Foshee2009). Thus, there is strong evidence that self-control can inhibit aggression and violence.
Recent research that has examined nearly all variables included in our model (Fig. 1) – measures of life history strategy, time orientation, self-control, and aggression – found that longer life expectancy is associated with an increase in the willingness to engage in behaviors reflective of a slow life strategy, whereas shorter life expectancy is associated with an increase in the willingness to engage in behaviors reflective of a fast life strategy (Dunkel & Mathes Reference Dunkel and Mathes2011). Shorter life expectancy is also related to short-term mating and less self-control, including greater willingness to engage in aggression, sexual coercion, and violent criminal acts, whereas the opposite is observed for longer life expectancy (Dunkel & Mathes Reference Dunkel and Mathes2011; Dunkel et al. Reference Dunkel, Mathes and Decker2010a; Reference Dunkel, Mathes and Papini2010b). When facing environmental uncertainty, individuals adopt a present orientation that is reflected in a fast life strategy, which in turn leads to more risk taking in phenotypic strategies related to reproductive success, such as interpersonal aggression (Kruger et al. Reference Kruger, Reischl and Zimmerman2008). More generally, these findings are consistent with our larger claim that fast and slow life-history strategies are linked to time orientation and self-control, which are likely to inhibit aggressive and violent behavior. At the same time, we should note that future research should examine the mediating role of time orientation and self-control on aggression and violence.
3.3. Conclusions
Individuals developing in warmer climates, where there is little seasonal variation and the environment is harsh and unpredictable, tend to adopt faster life strategies, a stronger present orientation, and lower levels of self-control. In contrast, individuals developing in colder climates, where there is much seasonal variation and the environment is not as harsh and highly predictable, tend to adopt slower life history strategies, a stronger future orientation, and higher levels of self-control (see Fig. 1). These mechanisms are essential to the development of aggression and violence. We are not suggesting that orientation to the future and high levels of self-control serve to inhibit all forms of aggressive behavior or violence. Our CLASH model focuses on “hot,” impulsive, reactive aggression and violence, for which longer time orientation and self-control are especially relevant.
4. CLASH: The ubiquity of climate (and latitude)
CLASH is not the first model to emphasize the important role of climate in affecting human thought, affect, and behavior. Indeed, climate is increasingly considered a powerful determinant of human behavior across a variety of scientific disciplines, including biological and evolutionary sciences (e.g., Epstein Reference Epstein1999), economics (e.g., Burke et al. Reference Burke, Hsiang and Miguel2015), and psychology (e.g., Van de Vliert Reference Van de Vliert2013a). In these disciplines, several topics are now being studied (e.g., health, welfare, happiness). Likewise, the empirical relationship between higher temperatures and increased violence has been demonstrated in many settings. For example, a meta-analysis found substantial effects of temperature increases on the likelihood of interpersonal and intergroup conflict around the world (e.g., Burke et al. Reference Burke, Hsiang and Miguel2015): One standard deviation increase in temperature was associated with a 11.3% increase in intergroup conflict and a 2.1% increase in interpersonal conflict. Examples of the increase in interpersonal conflict include spikes in domestic violence in India and Australia, greater likelihood of assaults and murders in the United States and Tanzania, ethnic violence in Europe and South Asia, and civil conflicts throughout tropical countries. We conclude that differences in average temperature and differences in seasonal temperature variation both help explain cross-national differences in aggression and violence around the world.
As noted earlier, distance from the equator can be used as an approximation of higher temperatures and smaller seasonal variation in temperature. In adopting that proxy, note that the term equator is not only defined geographically (at 0° latitude). The meteorological equator is located north (at 6°N) and what has been termed the “biological equator” is even further north of the geographical equator (at 10°N [see Aschoff Reference Aschoff and Aschoff1981, p. 481]). The biological equator is characterized by maximal ground temperatures, converging winds, and maximal cloudiness and rainfall. Although it is logical to use the biological definition of the equator (because of maximal ground temperatures), it is not entirely clear whether the biological definition is superior in terms of seasonal variation in temperature. Future research should consider all three definitions of the equator (see Douglas & Rawles Reference Douglas and Rawles1999). Furthermore, although distance to the equator can serve as a proxy, more precise predictors would be average annual temperature and seasonal variation in temperature. These sources are readily available and would help test the predictive ability, validity, and generality of CLASH across life strategies, time orientation, and self-control, as well as aggression and violence.
From an evolutionary perspective, hot and cold climates have posed divergent problems to human survival, which have required distinct psycho-behavioral adaptations (Murray Reference Murray2013; Van de Vliert Reference Van de Vliert2013a; Van de Vliert & Tol Reference Van de Vliert and Tol2014). The adaptive problems posed by very hot and very cold climates vary in their immediacy. In colder regions further away from the equator, the major challenge is to “heat and eat” (e.g., Van de Vliert Reference Van de Vliert2009), each of which requires coordination and planning in terms of timely harvesting, production of goods, and maintenance of stock and supplies. In regions closer to the equator, people face different and often more immediate challenges. As noted earlier, in warmer climates, it is not only the heat itself that challenges survival, but also pathogens and predators (e.g., Epstein Reference Epstein1999; Fincher & Thornhill Reference Fincher and Thornhill2012; Schaller Reference Schaller2006; Thornhill & Fincher Reference Thornhill and Fincher2011). Pathogen stress is strongly related to distance from the equator because temperature is an important determinant of disease transmission. Closer to the equator, where there is less seasonal variability, there is no cold winter to kill the many viruses and bacteria native to these areas (Schaller & Murray Reference Schaller and Murray2008).
Cold winters kill not only viruses and bacteria, but also the hosts that transmit them, such as mosquitoes that transmit malaria to humans (Blanford et al. Reference Blanford, Blanford, Crane, Mann, Paaijmans, Scheiber and Thomas2013). Infectious diseases, which proliferate in hotter climates, are an important cue of environmental harshness and unpredictability because they have caused more deaths in traditional equatorial cultures than predators, natural disasters, and war combined (e.g., Gurven & Kaplan Reference Gurven and Kaplan2007; Inhorn & Brown Reference Inhorn and Brown1990). Because environments closer to the equator are characterized by harsher and more unpredictable conditions, Life History Theory predicts that the individuals living there place greater value on the present more because they have a lower life expectancy. Predation risk is another cue to environmental harshness and unpredictability prevalent in hot climates with less seasonal variation in temperature. In these warmer locations live more dangerous animals, especially venomous animals, that can lower life expectancy and motivate individuals to adopt a faster life strategy, along with a stronger present orientation and a weaker focus on self-control.
5. CLASH: Caveats and future directions
As a model, CLASH is parsimonious because it focuses on two climatological variables (average temperature and seasonal variation in temperature). It is also a general model because it generalizes across socioeconomic and political-historical variables. Of course, this is not to imply that socioeconomic and political-historical variables do not influence aggression and violence. Indeed, some socioeconomic and political-historical variables are themselves (strongly) influenced by climate, so that they become “bad controls” for cleanly examining the causes of aggression and violence (cf. “bad controls” [see Burke et al. Reference Burke, Hsiang and Miguel2015]). Nevertheless, we suggest the relevance of some key variables that might enter our CLASH model. We call them “other key variables” because it is logically impossible to assign them the exclusive status of moderating variables, as they may be influenced by climate as well. For example, it is possible that a variable (e.g., wealth) can both mediate and moderate the effect of variable x (e.g., seasonal variation in climate) on variable y (e.g., time orientation) (for a discussion, see Hayes, Reference Hayes2013).
5.1. What about the role of wealth?
Do all people respond and adapt in similar ways to climate differences? There is some evidence that people from lower social classes, who typically have fewer resources, are more likely to adopt a fast life history, including the desire to reproduce sooner. Consistent with Life History Theory, a key variable is whether people grew up in a resource-scarce or resource-rich environment (e.g., Griskevicius et al. Reference Griskevicius, Tybur, Delton and Robertson2011). It is possible for people who face great seasonal variation in temperature to adapt in ways that make them more resourceful by devoting effort to preparing for the future, individually and collectively. When seasonal variation in temperature is low, less preparation and planning are needed, which may result in less building for the future, individually and collectively. This reasoning is consistent with CLASH and suggests that over time, cultures have evolved such that economic growth and prosperity decrease as distance to the equator decreases. Also, this reasoning may help explain the existence of what has been described as the equatorial Grand Canyon, a hot belt several thousand kilometers around the equator, characterized by an exceptionally large concentration of lower-income countries (e.g., Landes Reference Landes1998; Parker Reference Parker2000). Thus, violence and poverty often operate in concert because they are both rooted in a fast life strategy and focus on the present and little self-control.
A complementary line of reasoning may be derived from the Climate-Economic Theory of Freedom (Van de Vliert Reference Van de Vliert2009; 2013a), which emphasizes the combination of demanding climates and monetary resources. One especially relevant prediction of this theory is that monetary resources matter more in demanding climates: The rich can cope well because of their resources and even come to view demanding climates as challenging, but to the poor, demanding climates pose a genuine threat to survival. There is good evidence for this model in other social domains. A longitudinal study involving 123 countries found that generalized trust in strangers is determined by climate, primarily among the wealthier countries (Robbins Reference Robbins2015). Another study involving 74 countries found that adults in increasingly demanding cold or hot climates value cooperative enculturation of children if their society is richer, but value egoistic enculturation if their society is poorer (Van de Vliert et al. Reference Van de Vliert, Van der Vegt and Janssen2009).
These findings are in line with the Climate-Economic Theory of Freedom (for more evidence, see Kong Reference Kong2013; Van de Vliert Reference Van de Vliert2011b; 2013a) and underline the link of climate and wealth to self-control and aggression. On the basis of this theory it might be predicted that traits such as fast life strategy, time orientation, and self-control evolve especially among those who have plentiful resources. Those with fewer resources are not only more likely to adopt a faster life strategy, a stronger present orientation, and less self-control, but also they may face more fiercer conflicts over resources that might trigger aggression and violence.
5.2. Linear or curvilinear?
A classic issue in research on temperature and violence is whether the relationship between temperature and violence is linear or curvilinear (e.g., Baron & Bell Reference Baron and Bell1975; Reference Baron and Bell1976; Bushman et al. Reference Bushman, Wang and Anderson2005; Hsiang et al. Reference Hsiang, Burke and Miguel2013). This debate is quite complex and becomes even more complex in CLASH, which focuses on both average temperature and seasonal variation in temperature. When the focus is only on temperature, there is support for a curvilinear relationship. There is considerably more aggression and violence in warm climates than in cold climates, yet there is somewhat less aggression and violence in hot climates, that is, climates with average annual temperatures that exceed 24°C (75.2°F), which often are located inland and very close to the equator (see Van de Vliert Reference Van de Vliert2013a; Van de Vliert et al. Reference Van de Vliert, Schwartz, Huismans, Hofstede and Daan1999). Although comparisons between countries reveal a curvilinear relationship, comparisons within countries reveal a linear relationship. For within-country comparisons, which is our primary focus, it is important to note that many countries do not exceed the inflection point of 24°C (e.g., 60 of the 136 countries examined by Van de Vliert et al. [1999]; see also Bushman et al. [2005]). As a case in point, for comparisons within the United States, or within European countries such as Italy, or even within Europe as a continent, the parsimonious linear model is preferred over more complex, curvilinear models. Of course, for countries where the annual temperature exceeds 24°C, a different model should be advanced (see Van de Vliert et al. Reference Van de Vliert, Schwartz, Huismans, Hofstede and Daan1999). We recognize that CLASH is challenged in specifying how seasonal variation in temperature helps account for any deviation from linearity in the relationship between temperature and violence at the global level. Also, it is possible that the picture may be complicated by variables other than average temperature and variation in temperature, for example, differences in elevation, rainfall, distance to the ocean or sea, wind patterns, and factors linked to these geographical and climate differences, such as possibilities for agriculture, population density, and economic opportunity. For example, agriculture challenges planning and self-control, and “harbor cities” also challenge planning, self-control and organizational skill. In our view, this challenge is both theoretical and empirical and, therefore, in need of future research.
5.3. Future tests of CLASH: Many roads to Rome
5.3.1. From distance to the equator to direct tests of CLASH
The extant literature on the link between temperature and aggression has emphasized distance from the equator as an important variable. Distance from the equator is only an approximation of average temperature and seasonal variation in temperature. Both temperature and seasonal variation are determined not only by distance from the equator, but also by distance from the ocean or sea. The smaller the distance from the ocean, the less continental the climate and, therefore, the smaller the variation in seasonal temperature. Oceans make the climate milder. Continental climates are characterized by very hot summers and very cold winters and may have the strongest influence on time orientation, resulting in slower life history, stronger focus on the future and self-control, and therefore lower levels of aggression and violence. Furthermore, differences in elevation are also linked to climate. Regions at higher elevations are characterized by extremely cold temperatures. Other variables correlated with climate (e.g., seasonal precipitation) also call for planning and self-control, but are less strongly related to aggression and violence (see Burke et al. Reference Burke, Hsiang and Miguel2015). Thus, other geographical variables determine climate and may be important variables in CLASH.
In most research, it is possible to adopt a bottom-up (“data-driven”) or a top-down (“theory-driven”) approach. The bottom-up approach seeks to include many variables as predictors (“causes”) and criterion or dependent variables (“effects”). The technique is a regression-analytic approach (or a variance-accounted-for approach [Batson et al. Reference Batson, Van Lange, Ahmad, Lishner, Hogg and Cooper2003]) that helps organize the predictor variables economically to optimize their joint ability to account for as much variance in aggression and violence as possible. This bottom-up approach allows one to view a broader picture of the world of aggression and violence and to derive precise estimates of the variance accounted for by various predictor variables (e.g., see average cold and heat demands for 232 countries [Van de Vliert Reference Van de Vliert2013a]). For example, through this bottom-up approach, we know that temperature, rather than rainfall, is the more important climatological determinant of aggression and violence. This bottom-up approach is exceptionally useful in testing CLASH.
We, however, recommend that the bottom-up (data-driven) approach be complemented by a top-down (theory-driven) approach. With a top-down approach, locations are preselected on the basis of average annual temperature and seasonal variation in temperature, with other variables controlled for (e.g., wealth, religiosity). This top-down approach has some empirical costs because only some countries that can be schematically organized in a 2 × 2 framework of temperature (high vs. low) and seasonal variation (high vs. low) can be compared, while controlling for other variables. This top-down approach allows illumination of the mechanisms (i.e., the mediating power of life strategy, time orientation, and self-control) underlying the presumed effects of climate on aggression and violence. Thus, there are “many (paradigmatic) roads to Rome” and many ways to deal with “bad controls” (i.e., variables that are plausibly by themselves influenced by climate [see Burke et al. Reference Burke, Hsiang and Miguel2015; Van de Vliert Reference Van de Vliert2013a]). In the final analysis, we recommend a combination of the bottom-up and top-down approaches to test key aspects of CLASH. Because most prior research has used the bottom-up approach, we emphasize the added value of the top-down approach.
5.3.2. What types of aggression and violence?
As noted earlier, CLASH focuses on reactive forms of aggression and violence that are due largely to poor self-control. CLASH seems especially relevant to the various forms of aggression and violence caused by “honor” threats (Nisbett Reference Nisbett1993; see also cultural masculinity, Van de Vliert et al. Reference Van de Vliert, Schwartz, Huismans, Hofstede and Daan1999). Informed by a recent meta-analysis, we also suggest that many forms of aggression and violence related to climate operate not only between individuals, but also especially between groups (Burke et al. Reference Burke, Hsiang and Miguel2015). Within psychology, there is strong evidence that aggression is easily activated between groups. Groups trust each other less than individuals, and they exhibit stronger tendencies toward mutual exploitation (e.g., Reinders Folmer et al. Reference Reinders Folmer, Klapwijk, De Cremer and Van Lange2012; Wildschut et al. Reference Wildschut, Pinter, Vevea, Insko and Schopler2003).
This raises several intriguing topics for future research. One topic is whether regional differences, religious differences, class differences, or other “cultural” differences yield conflict. If so, the name CLASH also applies to these situations (Markus & Conner Reference Markus and Conner2013). Another topic is whether distrust underlies many forms of climate-related aggression and violence. The reason is that people and especially groups may be more easily provoked if they immediately attribute a negative act to aggressive intent. Distrust breeds misunderstanding and conflict, which leads to situations that can activate climate-related aggression and violence. This line of reasoning is plausible because research has indicated that general trust in others is weaker in countries closer to the equator (e.g., Balliet & Van Lange Reference Balliet and Van Lange2013; Robbins Reference Robbins2015). Furthermore, we noted earlier that pathogen stress contributes to the harshness of countries closer to the equator (e.g, Guernier et al. Reference Guernier, Hochberg and Guégan2004). However, pathogen stress is also closely associated with tendencies toward collectivism, including tendencies to think and act to protect and serve the immediate social group rather than the entire collective (e.g., ethnocentrism [Fincher et al. Reference Fincher, Thornhill, Murray and Schaller2008]). A strong, prosocial orientation to one's own group often can be at conflict with other groups, especially when resources are scarce. For example, when deciding on the route of noisy planes, individuals and groups may lobby or protest in favor of their own community and seek rerouting of the planes so that they fly over communities other than their own. This line of reasoning also helps illuminate why climate is more strongly related to intergroup conflict than to interpersonal conflict.
6. Concluding remarks
Several useful theories have been proposed to explain differences in aggression and violence between those who live in warmer and colder parts of a country or the world. These include the General Aggression Model (Anderson & Bushman Reference Anderson and Bushman2002), Routine Activity Theory (Cohen & Felson Reference Cohen and Felson1979), and Culture of Honor Theory (Nisbett & Cohen Reference Nisbett and Cohen1996). The purpose of CLASH is not to replace these theories, but rather to offer another possible explanation of these relatively large differences in aggression and violence between and within countries around the world. CLASH focuses on differences in average temperature and seasonal variation in temperature as two key climate variables that account for differences in aggression and violence, and it reserves key roles for fast and slow life strategies, time orientation, and self-control.
CLASH helps account for differences in aggression and violence both within and between countries, regardless of the size of those countries. It is a society-level model that uses differences in the climate (a key aspect of the “physical” environment) as a starting point and then bridges psychological processes within individuals (emphasizing life strategy, time-orientation, and self-control) with social processes and outcomes at the level of groups, cultures, and societies. Most past theories of aggression and violence tended to focus on psychological process or societal differences. Thus, we believe that CLASH provides a logical and internally consistent theoretical framework that integrates psychological processes and societal differences that have evolved and ultimately are rooted in geographical locations that underlie strong differences in climate.
Although the merit of CLASH is primarily theoretical, we close by outlining some important implications for society. Assuming CLASH is accurate, it is interesting to consider that people's thoughts and behaviors may differ based on the physical circumstances their ancestors faced and that they themselves face. The world is getting smaller and smaller. Electronic and social media (e.g., WhatsApp, Twitter, Facebook, email) connect us to people all over the world. Yet people coming from differing ancestral histories and living in different locations face challenges of self-control in a variety of ways. A businessperson from London may expect a response the next day, but the alliance in Nairobi may want to take at least an extra day. If CLASH is correct, the same pattern should hold for within-country differences between a businessperson working in Chicago and the alliance working in New Orleans, or between a businessperson working in Melbourne and the alliance in Brisbane or Cairns (with London, Chicago, and Melbourne being relatively more remote from the equator and facing greater variation in climate). Although people may have an implicit or even explicit understanding of some cultural differences in time orientation and self-control, it is likely that such differences may contribute to misperceptions and misunderstandings in cross-national communication. This is important because a perceived lack of self-control may pose a serious threat to interpersonal trust, even in ongoing relationships.
The implications of cross-national communication processes are potentially far reaching, and may help illuminate challenges and problems in business transactions, in international negotiations about climate change, and even in many interactions between Northern Europeans and the refugees coming from various countries closer to the equator (e.g., Syria, Afghanistan, Somalia). Turning back to within-country variation, consider the regional differences in attitudes and communication styles even within such a (large) country as the United States (e.g., Andersen et al. Reference Andersen, Lustig and Andersen1990; Nisbett Reference Nisbett1993). Because communicating “honor” is especially important to people living in the South of the United States, it seems advisable to adopt a respectful style of communication for business and effective negotiation with individuals from these states. Reserving judgment and giving the benefit of the doubt is probably an effective mindset, because provocation may be more quickly elicited in individuals from Southern states than in individuals from Northern states and, once elicited, more quickly translate into aggression and perhaps even violence. According to CLASH, these differences are ultimately rooted in climate differences and therefore should be relevant to understanding important differences in aggression and violence among many countries around the world.
Target article
Aggression and violence around the world: A model of CLimate, Aggression, and Self-control in Humans (CLASH)
Related commentaries (28)
A climate of confusion
Aggression, predictability of the environment, and self-regulation: Reconciliation with animal research
An alternative interpretation of climate data: Intelligence
Bullying when it's hot? The CLASH model and climatic influences on bullying
CLASH's life history foundations
Climate is not a good candidate to account for variations in aggression and violence across space and time
Culture matters for life history trade-offs
Dimensions of environmental risk are unique theoretical constructs
Does distance from the equator predict self-control? Lessons from the Human Penguin Project
Hell on earth? Equatorial peaks of heat, poverty, and aggression
Inconsistent with the data: Support for the CLASH model depends on the wrong kind of latitude
More than just climate: Income inequality and sex ratio are better predictors of cross-cultural variations in aggression
Postcolonial geography confounds latitudinal trends in observed aggression and violence
Pragmatic prospection emphasizes utility of predicting rather than mere predictability
Reply to Van Lange et al.: Proximate and ultimate distinctions must be made to the CLASH model
Russian data refute the CLASH model
Sociocultural discourse in science: Flawed assumptions and bias in the CLASH model
Stuck in the heat or stuck in the hierarchy? Power relations explain regional variations in violence
The CLASH model in broader life history context
The CLASH model lacks evolutionary and archeological support
The importance of being explicit
The paradoxical effect of climate on time perspective considering resource accumulation
The role of adolescence in geographic variation in violent aggression
The role of climate in human aggression and violence: Towards a broader conception
Using foresight to prioritise the present
Warm coffee, sunny days, and prosocial behavior
Where the psychological adaptations hit the ecological road
Why the CLASH model is an unconvincing evolutionary theory of crime
Author response
The Logic of Climate and Culture: Evolutionary and Psychological Aspects of CLASH