Southern Hospitality
Our own research has shown that these patterns of conflict suppression and explosion can be found here in the U.S. South: That is, conflicts are often driven under the surface and then ultimately explode in an incident that might have been avoided if the offending behavior had been checked in time.
This notion is embodied in the old expression that “a southerner is courteous and friendly...until he is mad enough to kill” (Wilson 1989: 635). In a similar vein, folk wisdom holds that Appalachian men are “slower to anger or panic, but absolutely without reason or mercy once the fat was in the fire” (Thompson 1966: 172).Honor and violence in the South. Our previous work using attitude surveys and lab experiments has shown that white U.S. southerners hold to a version of a culture of honor stance, in which insults, threats, and affronts must be punished with violence (Nis- bett and Cohen 1996). Historically this has been true, and it is still true today (Fischer 1989; McWhiney 1988).
In opinion surveys, contemporary white southerners are far more likely than their northern counterparts to endorse one person hitting, and in some extreme cases, shooting, another person for a variety of offenses—from insulting words to physical affronts or attacks (Cohen and Nisbett 1994). And in lab studies, we have shown that white male southerners act out these aggressive tendencies when provoked. In a series of three experiments, we invited subjects to the lab and had them insulted by a confederate who was actually working with us. Subjects were walking down a long narrow hallway when our confederate, who was working at a file cabinet, slammed the file drawer shut, rudely bumped into the subject, and then called him an “asshole” after the collision.
The effect of this insult was profoundly different for southerners and northerners in our study.
In response to the insult, southerners became far more angry, whereas northerners tended to be far more amused. Southerners also became more cognitively primed for aggression after the insult (as shown by their more violent completion of a written scenario), and they acted out their aggressive feelings in subsequent meetings with others and in physical challenge situations. Perhaps most surprisingly, southerners responded to the insult physiologically in a different way than northerners did. Whereas northerners seemed to be unaffected by the insult, southerners showed dramatic increases in their levels of cortisol (a hormone associated with stress and arousal) and testosterone (a hormone associated with aggressiveness and competition) after being insulted (Cohen, Nis- bett, Bowdle, and Schwarz 1996).However, these lab experiments tell us only about one type of conflict—that is, a conflict that begins with a single direct and unambiguous insult. In a more recent study, we examined the process by which politeness norms and aggression norms work together and feed off each other as conflicts escalate over time. In this study, we wanted to expose white male northerners and southerners to a series of irritations and annoyances from another person and see how they dealt with the conflict as it progressed.
Conflict over time. We had northern and southern subjects come to the lab and participate in what they thought was a simulated “art therapy” session. For this session, they were to draw a series of pictures from their childhood using crayons, ostensibly to promote relaxation and get “in touch with their inner child.” As they were doing this, however, another subject (actually a confederate of ours) began to annoy the subject while the two were left alone.
The provocations from the confederate were all mild (unlike like the unequivocal “asshole” insult above), though definitely meant to irritate. For example, the confederate would repeatedly wad up his drawings and throw them away, missing the trash can and hitting the subject instead.
Or he would repeatedly call the subject “Slick.” Or he would steal the subject's crayons. Or he would comment negatively on the subject's drawings. The question was: how would northerners and southerners handle this situation?Relevant predictions derive from the different cultural styles of the regions. Northern culture does not have the South's emphasis on graciousness and politeness. Indeed, the reaction many southerners have to the North is that it is exceptionally rude. (“Conversation in New York, Roy Reed has written, is ‘hurled stones'” (Reed 1986: 70)). And conversely, the reaction many northerners have to the South is that it is claustrophobic and “phony” (see De Lisser 1996; Nethaway 1996). However, even if it's not at this extreme level, many will agree that in manners lies “one great remaining (North-South) difference” (Reed 1986: 68-69).
In our experiment, northerners should do more to head off a conflict in the beginning and be more up front about their hostility at first, but then they should eventually level off, never letting their anger go completely “over the top.” Southerners, on the other hand, should play it cool, staying polite and expressing little emotion; and then when some critical point has been reached, they should be more likely to express intense anger or explode.
Consistent with predictions, people from the two regions did handle our confederate differently. Northerners were more likely to respond to the provocation in the beginning with expressions of hostility and more likely to be confrontational (as rated by our confederate and by an experimenter who watched on a video feed in another room). However, when they probably realized their warnings did little or no good, they leveled out. About halfway through the study, northern anger and conflict behavior plateaued.
Southerners showed a distinctly different response. Through the first part of the experiment, they showed little or no reaction to the annoying confederate, betraying no hostile affect.
At some critical point, however, things changed. About half way through the experiment, southern anger and conflict behavior ratcheted up quickly and dramatically. Though they began the study absorbing the annoyances stoically, they ended the experiment with expressions of anger and conflict that were far more intense, unpredictable, and hostile than northerners had ever shown. These patterns were born out in statistical tests that examined how people from the North vs. South differed in the absolute levels of anger shown over time, the rate of change in anger shown over time, and the maximum upward jump in anger and conflict behavior from one point in time to another (Cohen and Vandello 1997).Politeness and homicide rates. In addition to the lab study data, we have some correlational evidence suggesting that this process of anger suppression and explosion occurs in the real world as well. The evidence derives from the following logic: If it is true that in a culture of honor, politeness norms are going to drive conflicts underground only to have them explode later, then it should be the case that where southern society is more polite, there will also be more incidents of lethal violence. For places without a culture of honor (in this case, the North), this should not be the case.
To test this prediction, ideally we would have data on the politeness of various cities. Unfortunately, such a data set does not exist, but there is something close. Levine and colleagues have collected data on how friendly and helpful various cities are. They looked at such things as United Way contributions per capita as well as examined how helpful people were when experimenters dropped pens on the street, requested change for a quarter, and posed as blind people needing aid crossing the street. They found, not surprisingly, that the South was the most helpful region of the country (Levine, Martinez, Brase, and Sorensen 1994).
Perhaps more interesting for the present purposes, however, is what helpfulness was associated with in the southern vs.
non-southern regions of the United States. We merged Levine's data with data about white male homicide offenders, age 15 to 39, in the corresponding counties. As one might expect, in the North, the friendlier a place was, the less lethal argument- and brawl-related violence it had, whereas in the South, the effect was slightly in the opposite direction. In the South, more polite places actually had slightly more argument- and brawl-related homicides (Cohen et al. 1997). The statistical interaction held when we controlled for factors such as income, size of the metro area, and percent non-Latino white. Importantly, this pattern was true only for argument-and brawl- related violence (which are likely to center on insults, conflicts, and affronts and thus involve the processes of anger suppression and explosion described above). The interaction did not hold for felony-related homicides, that is, homicides committed in the context of another felony such as robbery or burglary. In both the South and the Nonsouth, more helpfulness and friendliness was associated with less felony-related homicide.