Violence is not one behavioral pattern but several. The multifaceted and complex nature of violence has led to a number of proposed guidelines and classification schemes for studying its component parts. Behavioral scientists have worked to develop classifications by grouping together meaningful categories of violence that share common characteristics related to etiology and function. One approach has been to classify violence according to the underlying motivation of the aggressor. A frequently used distinction is between hostile and instrumental motivation. In hostile violence, the major goal is to inflict harm or injury. In other words, hurting is an end in itself. In instrumental violence, actions may cause harm but are not motivated by the desire to cause harm per se. Rather, they are motivated by goals such as taking resources from others. In both cases, this distinction depends on the individual's intent, not on the act itself.
Although not conceptually clean, this distinction has proved useful. Certain types of violence such as armed robbery, murder-for-hire, and terrorism generally are well planned, goal-directed, instrumental actions. Offenders are acting to maximize their benefits and minimize their costs. Many prominent models of criminal behavior emphasize the rational choice component of crime (e.g., Cornish and Clarke). This type of planned behavior is distinguished from more impulsive and hostile violent actions often characterized by loss of control, irrationality, and rage. Such impulsive violent behaviors are frequently labeled emotional violence and are linked with emotions such as anger and fear. Biological models of violence have identified distinct neural patterns that characterize each type of violence. For example, the "low-arousal" aggressor more likely to commit instrumental violence is underreactive and responds sluggishly to stressors. In contrast, the "high-arousal" aggressor who is more prone to hostile violence tends to be hypervigiliant and easily frustrated (Niehoff).
Another distinction between classes of violence that bears some similarity to the hostile/instrumental classification is the difference between defensive and offensive violence. This distinction has been fundamental to animal studies of aggression, with defensive and offensive aggression linked to stimulation of different areas of the brain. In humans, instrumental aggression is roughly analogous to predatory aggression although it is limited to intraspecies behavior. In other words, when humans kill animals for food it is generally not considered offensive violence in the same sense as killing a rival gang member. Similarly, emotional or hostile aggression in humans could be considered the analogue of defensive aggression in response to a threat or perceived threat. Studies of children have found differences in propensity for proactive aggression and reactive aggression, although some children score high on both types of aggression (Dodge and Coie). This work provides some empirical support for distinguishing between offensive violence that is unprovoked and defensive violence that is a reaction to another's provocation.
Clearly, different classification schemes serve different purposes. In everyday usage, violence is often divided into distinct classes based on criteria useful for description, dialogue, and public policy. Violence can be grouped into categories based on variables such as the agents of violence (e.g., gangs, youth, collective groups), the victims of violence (e.g., women, children, minority groups), the relationship between aggressor and victim (e.g., interpersonal, nonrelated), perceived causality (e.g., psychopathological, situational, learned), and type of harm (e.g., physical, psychological, sexual). These criteria are frequently combined to examine particular forms of violence, such as psychological abuse of women in intimate relationships, youth sexual violence, instrumental collective violence, and so on.
Some efforts have focused on developing classification systems that can guide prevention, intervention, and control efforts. Tolan and Guerra describe four types of youth violence: situational, relationship, predatory, and psychopathological. This is not an exclusive classification proposed to cover all types of violence, but rather provides some conceptual organization for structuring efforts to prevent or reduce violence. Each distinct type of violence is associated with different causal mechanisms and warrants a different type of intervention. For example, relationship violence is influenced more by anger and conflict than predatory acts of violence such as armed robbery of a stranger. Consequently, biochemical interventions that block anger arousal or conflict-resolution training programs that teach anger management skills may have some influence on relationship violence but much less influence on predatory violence.