One of the aspects of animal behavior that most fascinates me is how groups function. How does a collection of individuals coherently make decisions that affect the fitness of everyone in the group? Consider a swarm of honeybees. Each bee is completely reliant on the hive; a worker can't ditch the hive if things don't go the way that individual likes. Decisions like where to build a new hive are critically important because the group can't afford to split if everyone doesn't agree. I stumbled across a great review article a while ago (available here) that talks about how animal groups of many taxa come to decisions that affect the whole group, and this blog post will be a summary of that article.
Article details
- Conradt L, Roper TJ. 2005. Consensus decision making in animals. TRENDS in Ecology and Evolution. 20: 449 - 456.
- Department of Biology and Environmental Science, University of Sussex
Very brief summary
Animals in social groups frequently need to arrive at decisions that affect the whole group. These decisions can be categorized as either consensus decisions, where members choose (e.g. by voting) between two mutually exclusive actions, and combined decisions, in which the group's behavior is the amalgamation of each individual's actions. Consensus decisions differ in 1. the degree of conflict of interest in the outcome among group members, and 2. whether communication is local or global.
Glossary
- combined decision - members of a group individually choose between multiple actions. The behavior of the group is the amalgamation of those decisions- e.g. the movement of a fish school where individuals are constantly joining and leaving the school, the success or failure of a human department store based on sales
- consensus cost - the cost (food, mating opportunities, etc.) of adhering to the group's consensus decision as opposed to the behavior that would be optimal for the individual at that given moment
- e.g. leaving with the group from a foraging patch while you're still hungry
- consensus decision - members of a group choose between multiple outcomes with the intent of reaching a consensus
- e.g. where to build a new hive, which prey to go after in cooperatively hunting species, the flight path of migrating birds
- global communication - all group members can communicate directly with all other group members. This occurs in small groups (e.g. primate troops)
- local communication - group members can only communicate with their neighbors. This occurs in large groups (e.g. ant colonies, ungulate herds)
Article summary-----------------------------------------------------------------------------------
Human societies depend on consensus decisions to survive, from large-scale international agreements to small gatherings of a few people. Some of the most pressing problems today stem from failures to reach a consensus, such as inability to draft effective climate change legislation, negotiations over maintenance of nuclear weapons and sustainable water usage, and pursuing alternatives to oil and gas. Studying how animals reach consensus decisions may yield insights into why humans sometimes fail at them.
The fact that animals communicate non-verbally raises interesting questions about who makes decisions that affect the whole group and what happens when outcomes differ in how beneficial they are for different members of the group (i.e. conflicts of interest).
Who makes the decisions?
While researchers have often assumed that the dominant individual in a group leads consensus decisions (e.g. the alpha male deciding the next foraging patch), leadership is actually variable and there is no correlation between leadership and dominance. Species differ in the number of contributors in a decision, with a continuum ranging from unshared (very few contributors), intermediately-shared (e.g. one demographic, such as adult males), and equally shared.
It might seem that the 'smartest' individuals should be the ones making the decisions, but surprisingly accurate decisions can emerge when everyone in the group contributes. Say there are two foraging patches to choose from, and one contains a predator. Each group member has some idea of which patch has the predator, but no one's really sure. If a dominant male is 75% sure, there's a 25% chance everyone will get eaten if he's wrong. But if that male incorporates the decisions of others in the group (who, for simplicity, also have a 75% chance of choosing the safe patch), the probability of making a mistake shrinks to 16% with three members, 10% with five opinions, 7% with seven votes, etc. because the majority of the group would have to be wrong for the group to choose the predator patch. So, even if the dominant has a better idea of the right decision, its individual error is still larger than the combined error of inexperienced group members.
(Cool side note: this type of information pooling is well-known in humans. One famous example occurred at a cattle fair in 1906, when the statistician Francis Galton asked people to estimate the weight of an ox. Of the 800 responses, which varied tremendously, the mean was off by only 0.5 kg. Basically, all of the crowds' errors cancelled one another out, leaving an estimate more accurate than any individual guess.)
In small groups, it's possible for members to vote, such as by specific vocalizations, ritualized signals, body orientation, and initiation movements. Decisions follow the majority of votes, but little is known about how animals estimate the relative number of votes. In large groups, self-organizing rules could lead to equally shared consensus decisions when only local communication is possible. For example, simple rules like 'always forage when your resources drop below a threshold or when others are foraging' or 'move in the average direction of your neighbors if the difference is small or the direction of the majority if the difference is large' can dictate the behavior of large groups.
How much conflict of interest?
In many taxa, there is little conflict between group members over what to do because the goal is similar for everyone. In migrating birds, for example, everyone has more or less the same destination. In ants and bees, there is no individual benefit in arguing over where a new nest should be built. Decisions like choosing which deer to kill in cooperatively-hunting animals like wolves, or where a primate troop should rest after foraging, also require relatively little conflict of interest.
.jpg)
However, following what the rest of the group does frequently requires an individual to act sub-optimally. A simple example would be if the group is deciding whether to go to a foraging patch or a drinking hole; it's very unlikely that every individual in the group has the same relative
level of hunger and thirst, so some would benefit from eating while
others would rather drink. The preferences can extend to the type of
food itself; in groups of white-nosed coatis (left), for example, some members are
better at exploiting one fruit source while others are better at exploiting another, so the choice of which foraging patch to go to is significant. These costs are especially relevant when you consider that decisions about travel destinations have to be made several times a day, every day.
While it might seem best to leave the group and just do what's best for yourself, the anti-predator benefits of group living (e.g. spotting predators from further away, and reducing risk of being eaten by diluting yourself amongst others) outweigh the huge costs of leaving the group. Predators often have a marked preference for stragglers, for example. Hence, it's usually better to compromise and stick with the group.
One way many mammalian species have evolved to best deal with consensus costs is by forming single-sex groups. Males and females often have quite different metabolic demands, with females investing in offspring through pregnancy, lactation, and raising the young while males compete with one another for access to the females. Also, as the relative energy requirements decrease with increasing body size, allowing large herbivores to subsist on lower-quality diets than small herbivores (the Jarman-Bell principle), the larger the difference in body size between males and females, the more likely the groups are to sexually segregate. Red deer (right) are an example of a sexually dimorphic ungulate species that has intersexual social segregation.
Who makes the decisions?
While researchers have often assumed that the dominant individual in a group leads consensus decisions (e.g. the alpha male deciding the next foraging patch), leadership is actually variable and there is no correlation between leadership and dominance. Species differ in the number of contributors in a decision, with a continuum ranging from unshared (very few contributors), intermediately-shared (e.g. one demographic, such as adult males), and equally shared.
It might seem that the 'smartest' individuals should be the ones making the decisions, but surprisingly accurate decisions can emerge when everyone in the group contributes. Say there are two foraging patches to choose from, and one contains a predator. Each group member has some idea of which patch has the predator, but no one's really sure. If a dominant male is 75% sure, there's a 25% chance everyone will get eaten if he's wrong. But if that male incorporates the decisions of others in the group (who, for simplicity, also have a 75% chance of choosing the safe patch), the probability of making a mistake shrinks to 16% with three members, 10% with five opinions, 7% with seven votes, etc. because the majority of the group would have to be wrong for the group to choose the predator patch. So, even if the dominant has a better idea of the right decision, its individual error is still larger than the combined error of inexperienced group members.
(Cool side note: this type of information pooling is well-known in humans. One famous example occurred at a cattle fair in 1906, when the statistician Francis Galton asked people to estimate the weight of an ox. Of the 800 responses, which varied tremendously, the mean was off by only 0.5 kg. Basically, all of the crowds' errors cancelled one another out, leaving an estimate more accurate than any individual guess.)
In small groups, it's possible for members to vote, such as by specific vocalizations, ritualized signals, body orientation, and initiation movements. Decisions follow the majority of votes, but little is known about how animals estimate the relative number of votes. In large groups, self-organizing rules could lead to equally shared consensus decisions when only local communication is possible. For example, simple rules like 'always forage when your resources drop below a threshold or when others are foraging' or 'move in the average direction of your neighbors if the difference is small or the direction of the majority if the difference is large' can dictate the behavior of large groups.How much conflict of interest?
In many taxa, there is little conflict between group members over what to do because the goal is similar for everyone. In migrating birds, for example, everyone has more or less the same destination. In ants and bees, there is no individual benefit in arguing over where a new nest should be built. Decisions like choosing which deer to kill in cooperatively-hunting animals like wolves, or where a primate troop should rest after foraging, also require relatively little conflict of interest.
However, following what the rest of the group does frequently requires an individual to act sub-optimally. A simple example would be if the group is deciding whether to go to a foraging patch or a drinking hole; it's very unlikely that every individual in the group has the same relative
level of hunger and thirst, so some would benefit from eating while
others would rather drink. The preferences can extend to the type of
food itself; in groups of white-nosed coatis (left), for example, some members are
better at exploiting one fruit source while others are better at exploiting another, so the choice of which foraging patch to go to is significant. These costs are especially relevant when you consider that decisions about travel destinations have to be made several times a day, every day.While it might seem best to leave the group and just do what's best for yourself, the anti-predator benefits of group living (e.g. spotting predators from further away, and reducing risk of being eaten by diluting yourself amongst others) outweigh the huge costs of leaving the group. Predators often have a marked preference for stragglers, for example. Hence, it's usually better to compromise and stick with the group.
One way many mammalian species have evolved to best deal with consensus costs is by forming single-sex groups. Males and females often have quite different metabolic demands, with females investing in offspring through pregnancy, lactation, and raising the young while males compete with one another for access to the females. Also, as the relative energy requirements decrease with increasing body size, allowing large herbivores to subsist on lower-quality diets than small herbivores (the Jarman-Bell principle), the larger the difference in body size between males and females, the more likely the groups are to sexually segregate. Red deer (right) are an example of a sexually dimorphic ungulate species that has intersexual social segregation.Conclusions and Future Directions
Empirical evidence and theory suggest that consensus decision making is common in a wide variety of behavioral contexts and animal taxa, including humans. However, the mechanisms through which these decisions are made still require further research, as well as whether decision making is equally shared or unshared in small groups.
Information pooling via self-organizing rules (e.g. dispersing eusocial insects, homing or migrating animals) can confer fitness advantages not available to solitary decision makers. Put simply, living in a group gives you access to others' information, which can help you survive. However, to turn the statement on its head, little is known about how the advantages of being in a group affect the sociality of the species. For example, could the benefits of information pooling factored in the evolution of social group structure, or did something else drive the evolution of social groups (e.g. defense against predators) and the benefits of information pooling were peripheral? Examining information pooling in species with differing conflict of interest in consensus decisions may also help us better understand cooperation in decision-making.
The full text of this article is available here.
Image citations:
white-nosed coati: Wikipedia
red deer: animalpictures123.org
Consensus decision-making is a group decision-making process in which group members develop, and agree to support a decision in the best interest of the whole. Consensus may be defined professionally as an acceptable resolution, one that can be supported, even if not the "favourite" of each individual. It has its origin in the Latin word cōnsēnsus (agreement), which is from cōnsentiō meaning literally feel together.
ReplyDeleteConsensus decision-making