We are a very pro-social species

From Blueprint by Nicholas A. Christakis

In light of the paucity of data on antagonistic networks and given the possibly important role that negative ties might play in social structure, in 2013 my lab decided to tackle the topic in a comprehensive and large-scale manner. Villages in the developing world offer an especially appealing natural laboratory to evaluate the structure of negative ties because they are relatively closed social systems where people cannot easily avoid others whom they dislike. So, we mapped the social networks of 24,812 adults in 176 villages in a field site in western Honduras. Using our Trellis software, we asked respondents to identify “the people in this village with whom you do not get along well.”  I had wanted to make the question even more pointed—perhaps “Whom do you dislike?” or “Who are your enemies?”—but I was advised against this by my project manager, a local expert, who pointed out that Honduras had the highest annual murder rate in the world (86.1 homicides per 100,000 people) and respondents would likely not want to risk answering such questions.

We mapped each of the 176 villages separately and identified all the social connections, both positive and negative, within them, yielding the largest and most detailed study of antagonistic face-to-face ties ever conducted. The villages ranged in size from 42 to 512 adults. Using the three standard name generators, we found that people identified an average of 4.3 friends (which included kinship ties, such as a sibling or spouse), with a range from 0 to 29 (though most people had between 1 and 7 friends). A total of 2.4 percent of the people reported having no friends by these metrics.

The good news is that animosity was much less common than friendship. On average, people identified 0.7 other people they did not like (which could also include people who were kin).  A total of 65 percent of the people reported having no one they disliked. This was the case for 71 percent of the men and 61 percent of the women, meaning that women were either more fractious (perhaps because they also generally report deeper relationships) or (as my sister, Katrina, insists) they simply had better memories about prior social interactions. One particularly vexatious person identified sixteen other people she did not get along with. This woman, who lived in a village of three hundred and twelve people, had four people who did not like her. But on the flip side, she nominated eleven people as friends and thirteen nominated her.

Furthermore, seen from the opposite point of view, on average, people were disliked by 0.6 other people, and most people (64 percent) had no one who disliked them.  The most disliked person was a woman who lived in a village of one hundred and forty-nine people—she had twenty-five people who identified her as an enemy. She herself nominated four people as her friends and only two people as her enemies. Alas, just one person nominated her as a friend.

While the pattern of friendship ties was remarkably consistent across villages, animosity varied widely. Although the percentage of negative ties was 15.6 percent overall, this measure was only 1.1 percent in one village and as high as 40 percent at the other extreme. The environmental, social, and biological forces affecting the formation of positive ties generate greater conformity than those affecting negative ties. The current local environments and cultures of the various villages seemed to have shaped animosity much more than friendship. Maps of the resulting networks for four of our villages are shown in color plate 5.

Just as friends tended to reciprocate friendship, enemies tended to reciprocate animosity, but the rates of reciprocation were quite different: 34 percent and 5 percent, respectively. That is, if you named someone as a friend, it was likely that person had also named you as a friend, but if you named someone as an enemy, it was less likely that individual had also named you as an enemy. This difference highlights the fact that people have secret enemies more often than they have secret friends. People declare their friendships to each other but are less likely to state that they are enemies.

These detailed data gave us the chance to quantitatively explore certain old theories (and commonsense ideas) regarding social connections, namely, that these four principles should hold:

The friend of your friend is your friend.

The enemy of your friend is your enemy.

The friend of your enemy is your enemy.

The enemy of your enemy is your friend.

[snip]

For the first time using population-wide data at this scale, we were able to quantify that the friend of a friend was nearly four times more likely to be a friend than to be an enemy. Our data also confirmed the second and third rules listed above. However, we found no evidence for the fourth rule, which claims that the enemy of your enemy is your friend. This rule is logical, and it would seem to be true. But it is a bit of a delusion. The person who is one’s enemy’s enemy has a large chance of being a friend simply because friends are so much more numerous in any ordinary group. It’s like trying to figure out whether black birds are more numerous than red birds in one location compared to another. Since black birds are generally more numerous than red birds, we must take into account this background prevalence before concluding whether black birds are more numerous in this particular spot. Once the greater frequency of friendship in general is taken into account, it turns out that one’s enemy’s enemy is not more likely to be a friend. In fact, that person is actually more likely to be an enemy.

This appears to relate to the likelihood that both enemies and friends cluster together within communities in networks. In our analysis, we were able to allow for a third possibility—namely, that people could simply be strangers to one another. And people were much more likely to be enemies of people they knew than of people they did not know. One must know someone in order to be either a friend or an enemy. This simple observation means not only that the enemy-making process requires contact and familiarity, but also that enemies are more likely to be found within one’s own group than among people in other groups. We also found that the more friends a person had, the more enemies he or she was likely to have, with each ten extra friends being associated with one extra enemy. Overall, we found that similar underlying social processes, arising from repeated interactions, resulted in both positive and negative ties.

That is a lot of information buried in a lot of words.  These seemed the salient points to me:

Honduras is saturated with endemic violence with 86.1 homicides per 100,000 people

People identified an average of 4.3 friends (which included kinship ties, such as a sibling or spouse), 

The range of number of friendships was from 0 to 29 (though most people had between 1 and 7 friends). 

A total of 2.4 percent of the people reported having no friends by these metrics.

Friendship is much more common than animosity.

On average, people identified 0.7 other people they did not like (which could also include people who were kin).

A total of 65 percent of the people reported having no one they disliked. 

71 percent of the men and 61 percent of the women reported having no one they disliked.

Women were either more fractious than men.

On average, people were disliked by 0.6 other people.

Most people (64 percent) had no one who disliked them. 

The prevalence of friendship ties is steady across villages.  

Level of animosity varies by village from a low of 1.1 percent negative ties to a high of 40 percent.

If you identify someone as a friend, there is a 34% chance they also consider you a friend.

If you identify someone as an enemy, there is a 5% chance they also consider you an enemy.

It is significantly true (4:1)

The friend of your friend is your friend.

The enemy of your friend is your enemy.

The friend of your enemy is your enemy.

The more friends a person has, the more enemies he or she is also likely to have.  Each additional ten extra friends is associated with gaining one extra enemy.

One must know someone in order to be either a friend or an enemy. 

People are much more likely to be enemies of people they know than of people they do not know.

I recognize that Honduras is not likely to be representative of all world communities but the dynamics are probably not dissimilar.  For example, I accept that Honduras in absolute terms is more violent and that other regions or countries might have less absolute violence.  However; I would guess that those other communities have a similar distribution of degrees of enmity (from 1.1% to 40%).

Here are some observations which the above data might support.  I recognize that these conclusions go far beyond the stated knowledge of the study.  

People are far more disposed towards friendship than enmity.  They have 4.3 friends on average and 0.7 enemies providing a friendship to enmity ratio of ~6:1.

People are strongly pro-social with 27 (65/2.7) times more people reporting having no enemies as people reporting having no friends.   

A large plurality (42%) have no enemies and are not seen as an enemy by anyone else (65 percent with no one they disliked times 64 percent with no one disliking them).

Friendship creation is steady across geographies and circumstance but enmity creation varies by as much as a factor of 36 (40/1.1).  Some communities are much more socially toxic than others.  

Unbalanced reciprocity of relationship is significant.  Only 34% of those assed as a friend share that assessment.  Only 5% of those assessed as an enemy share that assessment.  Either we are not aware of these imbalances or we are aware of them and don't consider that imbalance as important.  

Familiarity brings both new friendships and new enmities at a ratio of 10 to 1.  

This last point has some implications for multiculturalism.  Multiculturalism has the assumption that familiarity with one another reduces tensions.  In other words, we can overcome prejudice with familiarity.  What this data seems to suggest is that is not true.  Familiarity will bring increased friendships but also, at a much smaller rate, increased enmities.  All we are changing is tensions based on uninformed bias to negativity based on informed relationships.  Might be better but it is worth understanding.