A patterned set of relationships within a social group in which certain individuals consistently win contests over resources with certain others.

The concept of dominance hierarchy emerged in the early twentieth century, originally from studies of chickens (where the term “pecking order” originated) and was extended to many other social species through the mid-twentieth century. For decades, dominance hierarchies were thought to be the central organising feature of group-living mammals, with implications for everything from animal behaviour interpretation to handling and training practice.

Contemporary research has progressively moved away from simple linear dominance frameworks toward more nuanced models. In horses, what hierarchy exists is bilateral (defined between pairs of individuals rather than across the whole group) and context-dependent (the same individual may win contests over food but not over space, or vice versa) rather than linear. In wolves, the original alpha-pack model has been retracted by its main originator (L. David Mech) and replaced with a family-group model in which the “alpha” pair are simply the parents. In dogs, dominance frameworks have been substantially complicated by evidence that dog social behaviour is highly variable and context-dependent.

The shift in scientific understanding has not fully reached popular training culture. Many training traditions still operate from the older simple-hierarchy framework, with implications for handling practice that the contemporary research does not support. The popular framings of humans as “alpha” within a horse’s social structure, or as “pack leader” within a dog’s, derive from the older simple-hierarchy view and are not supported by current research.

Dominance relationships do exist in many social species. The contemporary understanding is that they are typically more nuanced, more context-dependent, and less rigid than the older simple-hierarchy view suggested. Resource holding potential, individual relationship history, current motivational state, and many other factors interact to determine outcomes of social contests.

For practical handling, the implication is that frameworks built on the simple dominance model often produce welfare concerns and behavioural problems that frameworks aligned with current research do not.