The major transitions in evolution (from molecules to cells, from prokaryotes to eukaryotes, from cells to organisms, from individuals to eusocial groups) as well as the formation of collective groups from independent units (human groups, migrating locusts, biofilms) required a reduction in conflict among formerly independent entities and an increase in cooperative interactions. Indeed, in many of the major transitions, the individual subunit loses its ability to function and survive outside of the cooperative body. However, it is clear that conflict remains, indicating that its suppression is not absolute. For example, there is evidence for conflict within cells (e.g., selfish genes, or between maternally and paternally inherited genes), among cells within an organism (e.g., cancerous cells), and among individuals in a group, even in eusocial groups (e.g., dominant reproductive and subordinate nonreproductive individuals).
Why does this conflict remain? Is it simply that the conflict is at low enough levels that it can import some selective advantage to the individual without negatively impacting the group? Does conflict arise only when “mistakes” are made, such as mutations in cancer cells that lead to unregulated growth? Or are there any scenarios in which conflict among individual subunits actually imparts some benefit to the group? Perhaps the ability to compete provides a certain degree of flexibility among individual subunits in the group, allowing individuals to assume different roles as needed and maximize the benefits of division of labor. Additionally, competition and conflict among group members may serve to ensure that cooperative be-
havior continues to provide a selective advantage. Competition is common with various species, including humans, and is even fostered within certain social organizations to promote productivity. Would scenarios in which conflict is advantageous only be possible at certain transitions, such as from individual to collective groups, but not among genes in a chromosome?
While most studies have focused on the factors that facilitate the evolution of social behavior from solitary behavior, the reverse is also possible and may be relatively frequent. Indeed, phylogenetic studies of eusocial evolution in bees have demonstrated that multiple lineages switched from eusocial to solitary behavior, or from eusocial to parasitic lifestyles. Thus, there appears be a great deal of flexibility, including reversibility, along the path from solitary to collective behaviors. What causes these reversals? It is simply an imbalance between conflict and cooperation within these lineages? If so, what causes these imbalances? In collective groups that form and dissipate more readily, are there dynamic changes in the levels of cooperation and conflict that mediate formation, maintenance, and breakdown of the group?
The challenge to the working group is to evaluate the degree to which cooperation and conflict need to be balanced in order to facilitate the evolution, expansion, optimal performance, and maintenance of collective behaviors. How does this balance vary depending on the nature of the units involved (e.g., cells, animals, and humans)?
Is it absolutely necessary for cooperation to be maximized and conflict minimized in order for collective behaviors to evolve? Or can high levels of cooperation support high levels of conflict?
Does the presence of conflict among individual subunits within a collective provide any selective advantage to the group, or is it always negative? Does it vary by species or units of analysis? How?
Is the same balance between cooperation and conflict required to facilitate the evolution, expansion, optimal performance, and maintenance of collective behaviors, or can/should this balance vary among these different stages?
Fraternal major transitions allow the cooperative benefits to flow to relatives while egalitarian ones necessarily must benefit directly. What are the ramifications of these differences? How does one inform the other?
What balance between cooperation and conflict is needed in order for collectives to reach the point of “no-return,” where reversion of the individual subunits to independent life is impossible?
What conditions facilitate the evolution of solitary behavior from social behavior?
What factors increase the level of cooperation among individuals in a group? Kinship is thought to be a major driver of social evolution, but some of the major transitions (molecules to cells, prokaryotes to eukaryotes) were likely not facilitated by kinship.
Danforth BN, Cardinal S, Praz C, Almeida EA, and Michez D. The impact of molecular data on our understanding of bee phylogeny and evolution. Annual Review of Entomology 2013;58:57-78.
Hughes W, Oldroyd B, Beekman M, and Ratnieks F. Ancestral monogamy shows kin selection is key to the evolution of eusociality. Science 2008;320(5880):1213-1216.
Maynard Smith J and Szathmary E. The Major Transitions in Evolution. Oxford University Press, New York, 1995.
Queller DC. Cooperators since life began. Review of: J. Maynard Smith and E. Szathmáry, The Major Transitions in Evolution. Quarterly Review of Biology 1997;72:184-188.
Queller DC and Strassmann JE. Beyond society: The evolution of organismality. Philosophical Transactions of the Royal Society B 2009;364:3143-3155.
Because of the popularity of this topic, two groups explored this subject. Please be sure to review the other write-up, which immediately follows this one.
IDR TEAM MEMBERS—GROUP A
J Arvid Ågren, University of Toronto
Danny Z. Chen, University of Notre Dame
Chelsey B. Coombs, University of Illinois at Urbana-Champaign
Margaret C. Crofoot, University of California, Davis
Rob Horsch, The Bill & Melinda Gates Foundation
Liliana B. Lettieri, Michigan State University
Jacob A. Moorad, University of Edinburgh
Corrie S. Moreau, Field Museum of Natural History
Adrien Querbes, Carnegie Mellon University
Srilakshmi Raj, Cornell University
Kelly A. Rooker, University of Tennessee, Knoxville
Chelsey B. Coombs, NAKFI Science Writing Scholar New York University
IDR Team 7A’s challenge was to focus on evaluating the degree to which cooperation and conflict need to be balanced to facilitate the evolution, expansion, optimal performance, and maintenance of collective behaviors. Collective behavior is defined as a spontaneous behavior that a large number of individuals within a group undertake that is different from the group’s previous actions. The classic example of collective behavior is seen within a honeybee colony, in which evolution selected for individuals to work together in various castes, from the worker bees to the nurse bees to the queen bee, and form societies that are truly well-oiled machines. Of course, collective behavior can be seen in human societies or groups of cells and a variety of other societies, as well.
The team began with a thorough discussion about the definitions of cooperation and conflict, which shaped all of Team 7A’s key questions and made up a significant portion of the topic. However, the team found that while cooperation is well defined, the definition of conflict is a bit murkier. It was agreed that cooperation is something that brings a net benefit to one or more players, while conflict occurs when fitness interests are not fully aligned among interacting entities.
While the fitness benefits of cooperation are often rather straightforward, the benefits of conflict may be less clear. Conflict arises when individuals, be they humans or honey bees, need to consume limited resources. Altough such conflicts may lead to tragedies of the commons and population extinctions, conflict may also play a productive role. To be productive, conflict does not have to be optimizing, it just has to be something that doesn’t completely destroy the group. Moreover, conflict between certain groups often heightens cooperation among those in specific groups. For example, in competition between businesses, one business might come up with an innovative strategy to fix a problem. Not only will the individuals in this innovative business work harder to implement this strategy, but so will the business’s competitors as they want to come up with their own strategy. Therefore, conflict can be beneficial at the group level.
The team decided to focus on a diagram by Washington University professors of biology Joan E. Strassmann and David C. Queller that represents conflict-cooperation space (Figure 1). It features two intersecting axes,
the x axis representing conflict and the y axis representing cooperation. In this plot, various groups like chimpanzees, coral, social aphids, and human cities are represented, and their interactions are qualitatively plotted, showing where the groups lie in terms of conflict and cooperation.
The team believed the “more conflict, more cooperation” quadrant seemed like the most interesting to delve into because it matches how we primarily characterize human-to-human interactions. The team also wanted to know what causes an individual or group to be unable to move back to a specific quadrant or place on the plot, or to be able to determine the point of no return. For example, honeybees, operating in eusocial societies, have evolved to a level with a lot of cooperation with very little conflict. However, if they cooperated even more or had even less conflict, would their eusocial societies fall apart?
There was a heavy focus on this point of no return. The team reasoned that the system may never be optimal, but a certain level of conflict and cooperation can help the system from breaking apart completely. A system breaking apart could look like the species or group dying out. After some
discussion, the group also decided that although, in theory, having the greatest amount of cooperation with no conflict might seem desirable, maintaining certain levels of conflict is good; if there is an environmental change, some individuals or groups will be better able to weather the change if conflict exists. This is natural selection exemplified. However, the team reasoned that there is no life at all when there is no conflict or cooperation.
Strassmann and Queller’s plot shows large clusters and gaps, and this spurred conversation as to whether these gaps are truly possible, which would suggest that certain combinations of conflict and cooperation make groups go extinct, or whether they are a function of the plot being made using only qualitative and no quantitative data.
After this lengthy discussion, Team 7A came up with its own statement of the problem:
Identify the modalities of combining cooperation and conflict to
1. Be beneficial to an individual or society—what are the successful combinations?
2. Determine dynamic pathways that lead to stable/improved states.
3. Predict a point of no return.
As a solution, the team thought it would be a good idea to expand upon Strassmann and Queller’s plot by adding a third dimension, or a z axis, that represents fitness (Figure 2). This z dimension would look like a fitness landscape, which is a 3-D plot that shows how well an organism or group does at reproducing. This z-axis various peaks and valleys would represent organisms, societies, groups, etc. and their relative levels of fitness. These peaks and valleys would be based upon the levels of conflict and cooperation inside an organism or group.
This new representation of Strassmann and Queller’s plot would show where phylogenetic pathways continually reoccur, and the number of species in certain parts of the plot would show at what levels of interaction there is more likely to be extinction or survival of species.
However, the problem with this 3-D representation is that although we are interested in the point of no return, with a fitness landscape, it looks as if the groups can always go back to where they came from originally. This is because in a fitness landscape, any individual at the top of a peak can fall back down into a valley, and is generally free to move around the landscape. This point of no return is therefore not accurately represented using this model.
To solve this problem, the group decided to create another model that projected the fitness landscape back into two dimensions with a grid (Figure 3). The nodes where the grid lines intersect represent groups or individuals, as well as where they reside in terms of their conflict and cooperation levels. There can then be pathways going between the various nodes to show where these groups or individuals could or could not possibly go; this effectively illustrates the point of no return in regards to whatever constraints lie in the individual or group’s particular environment.
With further experimentation, these particular models would be able to address whether there are areas on the plot where there are clusters or gaps among groups, as well as whether points of no return exist. The group also wanted to do further experiments to determine the role of scale in these models, the costs of moving from one place to another on the plot, or the costs of dealing with the interdependencies of entities as the social environment changes. As stated above, it is currently unknown whether humans are a true exception in that we have high levels of conflict and cooperation, and the group would want to do experiments to parse through that question.
To address the problem, the group recommended models like the NK model (fitness landscape), n-player game theory models, network models, and probabilistic models that would allow researchers to make predictions about various groups, including humans.
The group also suggested using comparative frameworks to identify
commonalities of the landscapes across scale. Scientists could also do experimental evolution using various organisms and microbes like green algae to increase the environmental complexity and modulate cooperation and conflict; this would allow them to determine whether there are points of no return or “dead zones” on the plot where species cannot exist.
From here, researchers could do simulations, and if enough quantitative information was derived, use big data strategies to solve the various facets of the problem statement above.
The team believes that the benefits to society these experiments will have can both promote the good and inhibit the bad. Our understanding of collective behavior and human cooperation can promote economic or other growth, be used to protect the environment, as well as support tenuous cooperation, like that of a ceasefire. We can also use points of no return to perturb cooperation in pests, pathogens, and disease, as well as prevent or stop war and restructure economic systems.
It will take a lot of cooperation—and some conflict—to better understand cooperation and conflict, but in the end, our society will be better for it.
IDR TEAM MEMBERS—GROUP B
Kevin E. Bassler, University of Houston
Sarah E. Evans, U.S. Department of State
William F. Fagan, University of Maryland
Alan J. Hurd, Los Alamos National Laboratory
Karim Nader, McGill University
Godfrey Onime, Johns Hopkins University
Dustin R. Rubenstein, Columbia University
Timothy E. Weninger, University of Notre Dame
Eva C. Wikberg, University of Tokyo
Godfrey Onime, NAKFI Science Writing Scholar Johns Hopkins University
IDR Team 7B was asked to evaluate the degree to which cooperation and conflict must be balanced in order to facilitate the evolution, expansion, optimal performance, and maintenance of collective behaviors.
Collective behaviors, or the coming together by members of a unit, can be immensely beneficial for the group. In humans, for example, the varied components of a cell—proteins, DNA, cell membrane—interact to form an efficient unit. Cells further work together in a complex individual, who in turn collaborates with other persons in various societal groups. These unified phenomena span different creatures and objects—from the collaborating parts of bacteria or humans, to atoms interacting to form intricate elements. But perhaps more intriguing, various individuals pool together to create complex groups—colonies of insects and bees, schools of fish, and, in humans, social groups, governments, and even armies. The units all work towards a common, if not always positive, purpose.
IDR Team 7B agreed that implicit in all collective behaviors is the need for cooperation, but conceded that conflict is also evident in most, if not all, interactions. The cancer cell goes rogue at the expense of the whole, competing queens fight for dominance in some species of social bees, dissenting parties are found in social media groups among humans, and murderers abound in virtually all human communities. The question for the team, then—to what extent cooperation and conflict affect the formation and stability of collective behaviors—seemed especially imperative.
Imperative, but complex. What was needed was to reduce the parts of the collective to the simplest units, while retaining enough complexity to make meaningful observations. As a first step, the team chose to limit its discussions to living organisms. It also considered the cells as the smallest unit of the collectives. To foster the interdisciplinary approach that the topic warranted, the team further allowed for, even actively solicited, the expertise of the physicists, computer scientists, psychologists, and policy analysts that complemented the biologists in the group. Equally important, the group appreciated the need for disagreements, simple or radical. With these ground rules, the team had become what it intended to study: its own interplay of cooperation and conflict in a collective unit.
The first challenge to answering an inquiry is to understand the question. The next, and perhaps most vital, step for the team became to eke out reasonable definitions of “cooperation” and “conflict.” An exercise of anonymous brainstorming recorded on yellow and pink and orange post-it notes quickly underscored the importance of the exercise. Initial thoughts on the meaning of cooperation ranged from “interacting individuals with shared interests,” through belief of what “increases happiness.” Other takes included “individuals working together for a specific outcome” and “parts working together for positive benefits for all or to lower free energy.”
After a spirited deliberation, the group settled on a working definition of cooperation as “interactions that increase utility function for the elements involved in the interactions.” There was a caveat: utility is context-dependent. In other words, the utility may differ depending on the group involved. For some eusocial insects, the context may be to protect the egg-laying queen and larva; for a terrorist group it may be to cripple a true or imagined foe; and for the atoms in an element, the utility function may be to achieve a state of order or to lower free energy.
Turning to conflict, the group reached a definition relatively more quickly: “Interactions that decrease the utility function of a subset of the elements involved in the interactions.” One caution was that conflict is not necessarily the opposite of cooperation. Conflicts must also not be viewed as “bad.” Such care to withhold moral judgments is important to view the interplay between the phenomena—cooperation and conflict—from a strictly utilitarian or functional perspective. The team reached a yet other consensus: Cooperation and conflict can occur simultaneously.
In research, areas of major or rapid changes offer the best chances for capturing the reasons for what is happening. The team needed to harness this power. Of the possible outcomes of the evolution, expansion, optimal
performance, and maintenance of collective behaviors, the power inherent in moments of change, or transitions, appears to center on when groups form and when they dissolve. The team opted to concentrate on these areas of transitions.
The ideal situation would be to study such periods of transition in evolution—when ants transitioned from being solitary individuals to being a formidable collective, or when some species of social bees revert back to solitary lives. But evolutionary phenomena rarely happen in human timescales. Hence, studying these transitions in nature would be less practical and entail huge cost, resources, and time—none of which the team had.
Fortunately, biology is replete with groups that form and stop rapidly, even cyclically: the murmur of starlings, tribes of bees, bacterial colonies, the cooperative breeding of some birds and animals, social collectives on the Internet, formation of nation states, and even phase transition in physics. Is there a utility function that can be generalized from these ephemeral groups, across scales, from cells to society? The team hypothesized that there likely is, but if that is not true, why? Further, what is the minimum number of phenomena that can describe the transitions—what are the simplest (but not too simple) utility functions at play? In the interplay between cooperation and conflict, what factors about temporal collective behaviors can researchers measure at the times they form and dissolve, factors that can be captured through an equation? The team conjectured that groups form when the utility function for the collective is maximized, and can only be achieved when cooperation exceeds conflict. When conflicts surpass cooperation, groups dissolve.
The next step in the inquiry would be to run the data gleaned from studying these varied examples in nature through a few or several models, to uncover what stands out. The results can then be plugged into more permanent collectively behaving groups in nature, to see what can be learned. Some promising models to start with could include
1. Cellular automata,
2. Agent-based models with adaptive/evolutionary dynamics, and
3. Formal population genetics model with probabilistic state transitions, which posits that states are phenotypes in cooperation × conflict space, and approaches like those from the neutral theory of biodiversity based on master equations.
Indeed, in the complex, interconnected world with an increasing con
straint on resources, it has become of dire importance to better understand the interplay between cooperation and conflict in the formation, expansion, performance, and maintenance of collective behavior. Such understanding is a necessary first step if we must control collective behavior. Perhaps then science can have a positive effect on human conflicts, endangered species, epidemiology and diseases, and—some members of the team insisted on adding—the U.S. Congress.