Deconstructing Duluth

Edward Dunning

The most popular hypothesis regarding the etiology of domestic violence is Power and Control; in relationships, a latent sense of male entitlement is manifest behaviorally when men attempt to exercise Power and Control over their female partners. As the philosophical keystone of the dominant model for explaining and treating domestic violence (Smith & Healey,1998), the construct of Power and Control is a central feature in contemporary analysis of domestic violence, and a required treatment component in most Batterers' Intervention Programs (BIPs): ninety percent of state certified programs "indicate that power and control issues are to be included in program content" (Austin & Dankwort,1997).

As illustrated by the Power and Control Wheel (Pence, 1988), acts which constitute domestic violence are divided into eight categories of behaviors:

Power and Control Wheel: Full graphic and text

Using Intimidation Using Coercion and Threats
Using Economic Abuse Using Emotional Abuse
Using Male Privilege Using Isolation
Using Children Minimizing, Denying and Blaming

Each category contains brief examples of behaviors which typically occur in discordant domestic relationships. As most couples who experience partner relational difficulties admit to engaging in some form of these behaviors, the Power and Control wheel provides a concise and accurate description of behaviors associated with domestic violence, and serves as an effective educational tool for assisting clients to identify and confront their own abusive behaviors. As descriptive model and educational tool, the Power and Control wheel is a powerful instrument for addressing phenomena associated with domestic violence. As an hypothesis, however, the original Power and Control construct is problematic.

First, the hypothesis is grounded in the notion of "generalized patriarchic oppression" as the root cause of domestic discord. Over time, a growing body of research has found little support for this "traditional, sociopolitical view[] of domestic violence" (Hamberger & Potente, 1996, p. 54). (Neidig & Friedman, 1984, McNeely & Robinson-Simpson, 1987, Vivian & Langhinrichsen-Rohling, 1996, Dutton,1996, Miller, S,1996, Moffitt & Caspi,1999, O'Leary & Murphy,1999).

Next is the problem of poor operational definitions: a prominent feature of the Power and Control hypothesis is indiscriminate use of the word "violence" to describe specific behaviors which do not involve actual violence. Although only two of eight behavioral categories, "Using Intimidation," and "Using Coercion and Threats," include behaviors which fit the existing definition of violence, all Power and Control Wheel behaviors are defined under the general category of "violence." First, operationalizing all inappropriate or abusive behaviors as "violence" confounds research and impedes comparison of studies (Tjaden & Thoennes, 2000). In practical application, partners are unfairly stereotyped and misdiagnosed when salient distinctions between behaviors are blurred, and "the confused misdemeanors inevitable in a relationship [become] stark crimes" (Young, 2000).

Two additional areas regarding practical application of the Power and Control hypothesis in BIPs warrant examination:

  • Strong sexual bias
  • The notion that Power and Control behaviors are purposive and premeditated.

In most therapeutic settings, formulating a diagnosis prior to meeting the client is not considered good practice. Yet- BIPs based on the concept of male Power and Control promote preconceived diagnoses based on gender stereotypes. The presumption that males have used violence to enforce authority, and females have used violence in self-defense is problematic in treatment of both male and female clients convicted of domestic violence related offenses. Whether overtly expressed or subtly implied, this "...tendency to depict women as perpetual victims and men as villains..." creates a treatment environment in which "Women's ill-treatment of men is either obliterated or excused" (Young, 2000). Attempting to fit the complex phenomenon of partner relational problems within this narrow perspective is not conducive to facilitating the healing process for male or female clients.

The negative effect of Power and Control on men

Most involuntary clients enter treatment to avoid incarceration (Mackenzie & Pendergrast,1992). Having entered a program, behavioral change alone is insufficient to maintain freedom; staying out of jail is dependent on demonstration of progress in group, and progress is typically measured by the degree to which clients express agreement with program philosophy. Male clients are required to address their partner relational difficulties in terms of patriarchic oppression; "Discussion centers on the actions used by the control his partner" (Smith & Healey,1998). Discussion of alternate hypotheses or theoretical models is not permitted (Florida, 1996, 2002), and clients who persist in exploring other variables, such as differences among individuals and their partners or the context in which domestic violence occurs, are pathologized as engaging in "minimizing, denying and blaming." Thus, regardless of the actual circumstances and under threat of incarceration, men are required to confess that their personal experience with "Domestic violence is... rooted in misogyny. It's about maintaining the position of 'king of the castle'" (Johnson, W., 2000).

In practical application, coercing male clients to embrace this rigid and self-pathologizing perspective under threat of incarceration "rob[s] them of dignity or...respect" (Corey, Corey, & Callanan,1993, p.340). However, many advocates and practitioners believe that humiliation of male clients is appropriate treatment in a program which "... should be seen not as an opportunity [for men] to improve themselves...It is in fact a punishment, a tool for holding them accountable." (Goodyear, 2001). First, focus on exacting accountability and retribution detracts from the object of eradicating domestic violence. Next, given the purported goal of protecting women, it is ironic that advocates have not recognized that a punishment-oriented approach may exacerbate the problem and put women at increased risk for serious abuse:

Humiliation; being stripped of one's dignity does something to the psyche that other emotions, such as shame, cannot... humiliation plays an important role in many murders" (corrosive effects, 1999).

Punishment-oriented, humiliating "treatment" correlates with high rates of client hostility, anger, rebellion and dropouts (Toseland & Rivas, p. 93, Smith & Healey,1998). Male clients who experience treatment as another link in a system of rejection are more likely to displace their frustration to partners: "One recent Florida study suggested that men were more likely to be rearrested after counseling" (Goodyear, 2001).

The negative effect of Power and Control on women

The concept of male Power and Control also inhibits treatment efficacy for women. Increasing numbers of women are arrested for domestic violence offenses, and researchers continue to find relative gender symmetry in domestic violence behaviors. (Straus, 1980, Gelles, 1987, AMA, 1994, Smith & Healey,1998, Moffitt & Caspi,1999, Fiebert, 2000, Tjaden & Thoennes, 2000). However, the dominant discourse on domestic violence continues to frame the problem in terms of mens' violence against women, and persists in ascribing etiology to patriarchic socialization of men: "women tend to be viewed as helpless, childlike victims...thus perpetuating conditions that may contribute to additional violence" (Neidig & Friedman, 1984, p.4). As this woman-as-victim hypothesis is incongruent with evidence that women are capable of perpetration, dissonance is resolved by explaining womens’ violence as a reaction to victimization: when women behave violently, they do so in self defense (Walker, 1989, Hamberger& Potente, 1996 Bachman, 1999).

In treatment of female clients, indiscriminate application of the self-defense hypothesis promotes focus on external locus of control and inhibits client development of insight. Like their male counterparts, the primary motivation of female involuntary clients is the desire to avoid incarceration. Staying out of jail is dependent on demonstration of progress in group, and progress is often measured by the degree to which clients express agreement with program philosophy. Clients are generally quick to figure out the demand characteristics of a program and learn to manipulate the personal bias of counselors. Promotion of the self-defense hypothesis by practitioners, whether overtly expressed or subtly implied, defeats the presumed treatment objective of eradicating domestic violence by introducing opportunity for offenders to blame their male partners and minimize their own behaviors.

The notion that Power and Control behaviors are purposive and premeditated        

Most certification standards have institutionalized the hypothesis that domestic violence is "The willful decision... to act violently" (Santa Clara, 1997). Although it is no longer considered a "best practice" for clinicians to ascribe motivation for client behaviors, counselors are enjoined to make it "...clear...that the use of...abusive behaviors were intentional acts designed to control their partners..." Florida Department Of Corrections (agency communication, 1996) in order to maintain certification. The words "intentional" and "designed" denote purposive and premeditated behaviors which are not related to interpersonal dynamics, emotional states, or external stressors, and strongly suggest the presence of sociopathic features. The blanket assumption that all clients fit this harsh and pathological diagnosis is not well-supported in domestic violence research. Almost twenty years ago, Neidig & Friedman (1984) developed the interpersonal perspective, which conceptualized this type of behavior as instrumental violence, operationalized in terms virtually synonymous with the concept of Power and Control: "...the deliberate use of violence as an instrument or tool for social influence" (p. 5). While acknowledging that instrumental violence does indeed occur in partner relations, the interpersonal perspective is based on a much broader framework which describes violence on continuum of expressive to instrumental. Expressive violence, the most common manifestation of domestic discord, is the result of emotionally charged interactions between partners:

Expressive violence expression or function of a high level of emotional arousal. It typically occurs in the context of gradually escalating conflict between [partners] (pp. 4,5)

Despite the fact that relatively consistent findings have, over time, demonstrated that this concept of partners as participants in expressive domestic violence accurately describes most cases, BIP counselors are charged with the task of interpreting all of the often unique and always individualized circumstances surrounding domestic violence as mens' use of instrumental violence. This narrowly defined perspective precludes discussion of any client's situation in terms which do not reinforce Power and Control philosophy. Discussion of alternative perspectives, or even the use of "... language suggesting [that] provocation, codependency, or mutual responsibility " might be factors is forbidden (Santa Clara, 1997). This means counselors are enjoined to ensure that clients frame their narratives solely in terms of instrumental violence. Thus in practical application, regardless of the actual circumstances, a male client's narrative must in some manner conform to the concept embodied in the statement:

"My wife doesn't do anything to upset me, but I slap her around just so she doesn't forget that I'm the boss."

The negative effect of Power and Control on practitioners

A significant problem in treating clients within this narrow framework is the multitude of client narratives, from men and women, which simply do not correlate with the hypothesis. An alternative, client-centered approach to understanding domestic violence, based on participant narrative, finds " informal questions allows the researcher to focus on the context in which the behavior occurs." Most often, participant self-reports describe a "continuum of provocation" in which both partners engage in mutually abusive behaviors which escalate to mutual physical aggression. Thus "The contribution the victim's role makes to the occurrence of violence cannot be evaluated in isolation from the other factors...with regards to violence in the family" (Gelles, 1987, pp. 29,157,164).        

A multitude of information describes this continuum of provocation, for example, the finding that "other wives tell us that they will actually provoke their husband to violence because they want him to be more dominant" (Gelles & Straus, 1988, p. 92) is relatively consistent with Laura Miller's observation regarding the intermittent violence in feminist icon Betty Friedan's 22 year marriage:

"I taunted him into finally beating me up" as a way of venting marital pressures. While some battering relationships are simple tyrannies, the reality is that others, like the Freidans' marriage, are more complex dances in which each party willfully torments the other (2000)

Despite the abundance of evidence, anecdotal and quasi-empirical, that mutual antagonism affects a significant number relationships, this "dyadic phenomenon" (Vivian & Langhinrichsen-Rohling, 1996) is dismissed by Duluth proponents as the "myth of provocation." This position is carefully protected from critical scrutiny by certification standards which prohibit exploration or discussion of mutual participation in domestic violence as “blaming the victim”:

Any treatment approach or practice that blames or intimidates the not appropriate...Any treatment provider who blames the victim is in violation of the principles of these STANDARDS (San Diego, 2001)

•Theories or methods which in any way bring the victim into the circle of responsibility for the batterer's behavior or diminish the batterer's responsibility for the violence are inappropriate. (Santa Clara, 1997)

•Systems theory approaches which treat the violence as a mutually circular process, blaming the victim are inappropriate (Florida,1996).

This puts counselors in a multiple bind; professional and ethical standards compel counselors to cultivate a relationship with each client in group, while certification standards restrict the nature of dialogue necessary to develop an accurate perspective of individual circumstances. Client narratives are often congruent with other models having at least equal empirical value, but as these models are "prohibited," "not allowed," or "inappropriate," practitioners must redirect clients to frame their subjective experience in pathological and blame-focused terms. Although in application this prohibits a “start where the client is” approach, control over therapeutic technique is justified as necessary in order to "eliminate the possibility that groups may become support groups to justify violence" (Santa Clara, 1997). Thus practitioners are prohibited from asking probative questions such as “'What triggered your behavior?' or 'How did things get so out of control.'" In any other therapeutic setting, such questions are accepted as standard practice. However, certification standards have determined that such questions “…provide[] rationalizations for abusive behavior.” Even expressions of accurate empathy, a classic client-centered technique, are considered to be "colluding comments" e.g., "I can understand how that would set you off," or "You shouldn't let her hook you into an argument" (Florida DOC).

The word "collude" appears frequently in Duluth-based criticism of alternative perspectives, denoting the perception that practitioners who use forbidden concepts or language, or in any manner question the accuracy of Duluth philosophy must be fraudulently conspiring with male clients to reinforce patriarchal oppression. One carry-over effect of the ideology of generalized patriarchic oppression is the perception that all men, including trained and professional practitioners, are influenced by the latent belief that males are entitled to dominate their female partners through physical force. As expressed by the executive director of a New York State certified program "that has been lauded by battered-women's advocates":       

I will tell you that there is not man working with me who I would say is or is not a former perpetrator...I don't ask. The reality is that men in this culture, by virtue of what we teach them, use controlling strategies and tactics with their partners as part of normal life. To weed out men who have ever done an abusive thing against a woman would mean there would be no instructors. (Goodyear, 2001)

Regarding that perception, and the frequently expressed accusation of counselor collusion, this statement must be strongly expressed:

  • Competent and ethical clinicians and researchers do not minimize, rationalize, justify, blame the victim, or in any other way collude with any form of violent, abusive, or oppressive behavior - irrespective of whether such behaviors are directed towards women, children, self, or any other minority.
  • Likewise, no competent and ethical researcher or clinician would accuse another of conspiring in fraud without hard evidence.

There is no question that individual thought and behavior has been influenced to some degree by patriarchy. There is no question some men do use instrumental violence and abuse to dominate women. The percentage of men who actually fit the criteria for intimate terrorism has not been well established, but some studies of male offender populations do not support the notion of generalized patriarchic oppression (Dunning, 2002). As more and better research becomes available, the notion that all, or even most incidents of domestic discord are related to instrumental abuse or violence becomes more difficult to support. At this time the weight of evidence indicates that most instances of domestic discord are related to the phenomenon of "common couple violence" (Johnson, M.P.,1995, O'Leary & Murphy, p. 30).

At this time, as no treatment method has yet demonstrated superior efficacy, there is no rational basis for promoting any model which precludes exploration and discussion of other variables in treatment. Consequently, Duluth-based certification standards have engendered harsh criticism from some "professionals in the mental health community," who find that "circumscribing practice in the absence of empirical evidence to support particular interventions is unfounded and indefensible" (Austin & Dankwort,1997).

Summary and Conclusion

As a descriptive model and educational tool, the Power and Control wheel is a powerful instrument for addressing behaviors associated with domestic violence. As an hypothesis, the original Power and Control construct offers a one-dimensional perspective which may be:

more or less 'correct' when addressing it's own level and grossly distorting when applied inappropriately to other bands of the spectrum (Cowley, 1996, p. 678).

Given these circumstances, the need for objective assessment of the of the full range of dynamics in partner relational difficulties and reevaluation of the mainstream approach to domestic violence is evident. The concept of Power and Control and the popular notion of all forms domestic violence as a dynamic of "batterer" and "victim" are two areas in particular which warrant further exploration.


© 2001, 2003

Author's note: In March, 2003 i received notification of a 1998 article by Dr Sally Satel, containing similar material:
The Patriarchy Made Me Do It. This is a must-read, reposted here with permission.