Dove

Dove is a hypothetical fighting strategy used to explain why natural selection can maintain a mixture of behavioral types rather than favoring either unrestrained aggression or nonviolence alone. Its defining restraint—threatening without causing injury—can be advantageous in one population and costly in another. The model’s central point is that a strategy cannot be judged in isolation: its payoff depends on what the rest of the population does.

Definition and Model

In the Hawk-and-Dove model, each individual is classified as either Hawk or Dove, and cannot identify an opponent’s type before fighting; it has no memory of previous encounters with particular rivals. Hawk fights as hard as possible and retreats only when seriously injured. Dove performs a conventional threat display but does not hurt its opponent. A Hawk facing a Dove wins when the Dove backs down; two Hawks continue until one is seriously injured or dead; two Doves posture until one tires or gives way. The label is a conventional name for a model strategy, not a description of the habits of real doves, which the book notes can be aggressive.

The model assigns 50 points for winning, zero for losing, minus 100 for serious injury, and minus 10 for time spent in a prolonged contest. These arbitrary values stand for effects on gene survival and make the comparison of average payoffs easier to follow. They are not meant as literal measures of animal behavior; the argument turns on how the cost of injury changes the relative returns of the strategies.

Payoffs and Population Balance

In a population made entirely of Doves, contests end without injury. Each individual wins about half its contests, receiving 50 points for a win but losing 10 to the delay; a loss also incurs the 10-point time penalty. The resulting average payoff is 15 points per contest. Dove therefore appears to do well when its opponents share its restraint.

That advantage attracts a contrasting strategy. A mutant Hawk among Doves meets only Doves and wins every encounter, earning 50 points per contest against their average of 15. Hawk can consequently spread. But if Hawks become common, they increasingly meet one another: each Hawk then wins half its fights for 50 points and loses half with a penalty of 100 for serious injury, giving it an average payoff of minus 25. A Dove among Hawks loses every contest but is not injured, so its average payoff is zero. In that population, Dove gains the advantage and can increase.

This reversal does not require an endless swing from one majority to the other. Under the example’s chosen point system, the stable population is seven-twelfths Hawks and five-twelfths Doves, or a Hawk-to-Dove ratio of 7:5. At that ratio, both strategies have the same average payoff, so selection favors neither over the other. If the Hawk share rises above the equilibrium, Doves gain an extra advantage and the balance moves back; fluctuations around the stable point need not be large. The equilibrium is not evidence that the population has arranged itself for the good of the group: it emerges from differences in individual payoffs as strategies interact.

Limits and Wider Context

Dove’s restrained contest is only one deliberately simplified model of conflict. The broader analysis asks how an animal’s decisions about whether to fight, escalate, or withdraw can be understood as cost-benefit calculations, without supposing that the animal consciously reasons through them. An evolutionarily stable strategy is one that cannot be improved upon by an alternative when most of the population already follows it. The Hawk-and-Dove example shows why the best response depends on the population’s existing composition, and why a stable outcome may combine strategies rather than consist of one universal winner.

The book’s later account of cooperation extends this emphasis on interaction structure without identifying Dove with cooperation in every setting. In the Prisoner’s Dilemma, players choose simultaneously between cooperation and defection, and the consequences depend on both choices. A divorce that appears to be a zero-sum dispute can, the book argues, instead be handled as a mutually beneficial nonzero-sum situation; professional rules that require spouses to use separate lawyers can help turn it into an adversarial contest. In the 1977 Bristol–Coventry match, news that a draw would keep both teams in the First Division changed the incentives during play: after Sunderland’s defeat became known, both teams had reason to secure a draw rather than keep attacking. These examples do not alter the Hawk-and-Dove payoffs. They reinforce the wider analytical lesson that behavior and outcome depend on the conditions and rules governing interaction, not merely on a label such as peaceful or aggressive.

Repeated encounters add another condition for cooperation. The account of the Repeated Prisoner’s Dilemma stresses that players must not know exactly when interaction will end: if both know the final round in advance, backward reasoning makes defection the predictable choice there and then in each preceding round. The Dove model instead assumes that individuals cannot tell an opponent’s type in advance and have no memory of specific past opponents. Read together, the examples distinguish a stable mixture of strategies under a particular payoff structure from cooperation sustained by the prospects and terms of future interaction.