The oft-used proverb “The enemy of my enemy is my friend” is not just limited to middle-school English classes and middlingly written revenge dramas. Not only does it spring from a socio-psychological theory proposed by psychologist Fritz Heider in the 1940s, it appears that it has finally been scientifically proven.
A research team from Northwestern University used physics and statistical theory to confirm Heider’s hypotheses. Senior study author Istvan Kovacs expands upon this much ado about “friends” and “enemies” and how his team arrived upon the conclusion that Heider, in fact, was right.
What is social balance theory?
The principal grounding of the social balance theory lies in the achievement of cognitive consistency, which is, in turn, a step toward psychological balance. Individuals strive for harmony in their interpersonal relationships, and when these relationships fall outside the ambit of the four rules described by Heider, this harmony is disturbed until the individual finds a way to restore it by altering how they feel about one of the constituent elements of the equation.
For example, if you like a celebrity that goes on to endorse a product that you dislike, your opinion of both the celebrity and the product would be undecided, hence leading to cognitive dissonance. That is, until you decide that the product isn’t so bad after all, or that you do not really like the celebrity. This achieves harmony again.
According to senior study author Istvan Kovacs, “Social balance can be considered in a group of people, usually among at least three people, forming a triangle. The connections between them represent the sentiment of relationships between each other, either positive or negative.”
“Balance appears if all relationships are positive or when two negative relationships are balanced by one positive relationship. Other situations lead to unbalanced triangles, causing tension or frustration,” he explained further.
Previous work on the theory
Despite numerous attempts by researchers to validate this theory using network science and mathematics, their endeavors have often not attained complete success due to the inherent complexities of social networks, which often deviate from perfectly balanced relationships (as posited by the theory).
Therefore, the crux of the matter lies in determining whether social networks exhibit a higher degree of balance than what would be expected according to a suitable network model. However, most existing network models have proven too simplistic to adequately capture the intricacies of human relationships, resulting in inconsistent findings.
The team from Northwestern University has managed to overcome this challenge by successfully integrating two crucial components essential for Heider’s social framework to function effectively. In real-life scenarios, not everyone is acquainted with each other, and individuals vary in their positivity towards others.
While researchers have long recognized the influence of both factors on social ties, existing models could only accommodate one factor at a time. By simultaneously incorporating both constraints, the resulting network model developed by the researchers has finally substantiated the much-studied theory nearly eight decades after Heider first proposed it.
Kovacs’s says, “As physicists, our initial interest was sparked by the 2021 Nobel Prize awarded to Giorgio Parisi who studies complex physical systems, shared with Syukuro Manabe, and Klaus Hasselmann working on climate modeling.”
He states that in Parisi’s studies, competing interactions between spins can lead to configurations unable to simultaneously minimize all pairwise interactions, resulting in inherent frustration.
Kovacs remarks that social systems sometimes exhibit analogous frustrations, where conflicting interpersonal relationships or attitudes can create a state of imbalance or tension within a social network.
“We found it exciting to combine insights and tools from statistical physics with large-scale datasets about real social systems,” he adds.
The research conducted by Kovacs’s team
To investigate the issue, Kovacs’s team turned its attention to four extensive signed network datasets previously assembled by social scientists—(1) user-rated comments on the social news site Slashdot; (2) exchanges among Congressional members on the house floor; (3) interactions among Bitcoin traders; and (4) product reviews from the consumer review site Epinions.
In their network model, the researchers departed from assigning purely random negative or positive values to the edges. In a truly random scenario, every node would have an equal likelihood of encountering one another. However, real-life social networks do not adhere to such uniformity, as not every individual is acquainted with everyone else in their network.
To enhance the realism of their model, they devised a statistical approach to allocate positive or negative values based on the likelihood of such interactions occurring. This method ensured randomness within the constraints imposed by the network’s topology. Furthermore, the team incorporated the notion that certain individuals are inherently more amiable than others, leading to a higher likelihood of positive interactions and fewer hostile encounters.
By integrating these two constraints, the resulting model consistently demonstrated alignment with Heider’s social balance theory across large-scale social networks. Moreover, the model revealed patterns extending beyond three nodes, indicating the applicability of social balance theory to larger graphlets involving four or more nodes.
Practical applications
This innovative framework holds immense potential in advancing our understanding of social dynamics, including phenomena like political polarization and international relations. Moreover, it can be applied to any system characterized by a combination of positive and negative interactions, such as neural networks or drug combinations.
One way to understand how this theory plays out in real, day-to-day life is, according to Kovacs, by “thinking about two options that people need to choose from (relevant in many settings, like political polarization): it would be ideal for friends to share the same opinion, and enemies to be on different sides. Such an ideal situation can only happen in balanced triangles.”
In accordance with social balance theory, people instinctively strive to alleviate tension to uphold social stability and comfort. This theory serves as a valuable framework for understanding our behaviors and motivations in establishing social interactions among individuals.
When it comes to applications within human neural networks, the senior study author says that neurons are connected through synapses, which can transmit either excitatory or inhibitory signals. This dynamic mirrors the interactions found in social networks.
“Our model could potentially leverage the known connectome data and any initial hypotheses about the excitatory or inhibitory nature of certain neurons to infer the likely characteristics of the remaining synaptic connections,” Kovacs explains.
Likewise, in the case of drug combinations, this model could offer valuable insights into whether the pairing of two drugs enhances efficacy in treating particular diseases.
While the social network study provided an ideal foundation for exploration, the authors posit that their primary focus is to extend their investigations beyond interactions among friends and delve into other intricate networks.
The study has been published in Science Advances.
