Leigh notes the powerful role that "dominance feminism" has played in an effort to craft a system that calls for arrest and prosecution in all cases, on the theory that women in domestic violence are controlled by violent partners and hence unable to act voluntarily. Leigh contrast this approach with "autonomy feminism" which focuses on women's expressed wishes. The abstract for the article states:
Domestic violence law and policy prioritizes the goals of policy makers and battered women's advocates - safety and batterer accountability - over the goals of individual women looking for ways to address the violence in their relationships. The shift of decisionmaking authority has profoundly negative implications for the autonomy of women who have been battered and reflects the influence of dominance feminism on the battered women's movement. This article argues that the time has come to shift the lens through which we view domestic violence law and policy from dominance feminism to anti-essentialist feminism,* allowing us to see how problematic mandatory policies are and helping us to craft domestic violence law and policy that honors the goals and priorities of women who have been battered.
The question isn't an easy one and doesn't deserve the clear resolution that the abstract suggests.
Part of the problem, in both domestic violence and incest cases, is that these breaches of a relationship frequently happen in the context of deep economic dependence. The short to medium term alternative to a relationship with domestic violence is often deep poverty. Courts are permitted to be more generous, but cannot be relied upon to provide for an economic dependent leaving a relationship.
Another part of the problem is that the psychological fear and need to submit. I've seen up close in my practice over the years the way that mental habits of fear and submission create what would not look like rational decision making to a disinterested third party, usually to the victim's detriment. It can take months or years before a domestic violence really understands her options and is genuinely acting autonomously. The fears are often enhanced and sustained by shared parenting arrangements that create options for strategic retaliation which is difficult to prove by one unscrupulous parent against co-parent. In some ways, insisting on the prosecution of domestic violence perpetrators is valid for the same reason that police intervene to stop suicides. The assumption is that in the vast majority of cases, the person helped will appreciate that decision in hindsight.
Then there is the issue of social responsibility, and saving face. If the police intervene to stop a domestic violence incident over the public objections of a domestic violence victim, the victim may escape the offender's wrath after the fact if the relationship continues, because the consequences of the act did not flow from a decision made by the victim (even if the intervention was privately appreciated). Even the mere possibility of an arrest and prosecution over the objection of a victim may discourage a potential domestic violence perpetrator from being overconfident that he will be able to escape the consequences of his act because he believes that he can control the victim.
There are also arguments in favor of a strong role for autonomy.
First, there is the risk of a mistake of fact by a third party about what is going on in an incident, particularly if it is an isolated first encounter. A law enforcement officer encountering a single incident in which there could be probable cause to believe that domestic violence is present, is often ill equipped to know if what would sound like a lame excuse for the cause of an injury in the context of repeated domestic violence calls really was simply a quirky accident. Enthusiastic sex, seen in a momentary glance before it is interrupted, can look very much like rape to an outsider who lacks of the context of those involved.
Likewise, while many relationships in which there is a violence incident do escalate into tragedy, some don't. Men who are tamed by their mates after a violent incident and go on to make good partners are not unknown to history. A woman may rationally decide to stay with an otherwise good enough man who slaps her, particularly if she thinks she understands why that happened well enough to manage the risk of that happening again in the future.
An autonomy approach also makes possible intermediate levels of law enforcement intervention. Ending law enforcement discretion forces women into an all or nothing choice to "go nuclear" with a call likely to end in an arrest, prosecution and conviction, or not call for help at all. Yet, in almost all other areas of law enforcement, it is much more common for a law enforcement officer to come to the scene and defuse an incident when trouble is brewing but no serious crime has been committed, than it is for a law enforcement officer to actually make an arrest. The assumptions about what police do in non-domestic cases to deal with low level disorderliness and minor violence which are implicit in mandatory arrest and prosecution policies aren't necessarily accurate. For someone who has a lot to lose because of a situation of dependency if the perpetrator goes through the criminal justice system and loss of professional reputation as a result, a mere show up by police to the scene may prevent a bad situation from getting out of hand without catastrophic consequences, may seem like a good option to have available.
And, sometimes domestic violence victims who look like they are being irrational, are simply looking at their enlightened long term best interests. Once the immediate incident is under control, a divorce and temporary restraining order can often be promptly obtained, ending the relationship legally with state supervision. An excessive focus on economic wins and losses in the wake of an ended relationship in a divorce may not be that important, relative to a swift end of the marriage, in the common situation where a remarriage follows swiftly and becomes a core of economic support, and a new relationship follows even more swiftly. The economic value of the new relationship may overwhelm what is to be gained or lost from a relationship that is already over. Preserving some sort of relationship for an extended period of co-parenting may be more important for a domestic violence victim than writing past injustices committed by a domestic violence perpetrator.
There is a middle ground between zero tolerance, which often also means zero common sense, and taking a victim incapable of acting voluntarily at her coerced word. Not all domestic violence incidents are hopeless, but many are predictable tragedies that the victims are incapable of taking seemingly obvious steps to prevent. In some cases, maybe many, arrests should be made, and prosecutions should go forth, over the objections of the victim. Finding sensible middle ground isn't easy, and shouldn't be, because the issues are inherently complex.
The real question is how to establish a policy that respects the intelligence of the law enforcement officials, prosecutors, judges, private lawyers and other individuals who make decisions in these cases at critical points, while suppressing what was, at one point in time anyway, entrenched biases towards domestic violence perpetrators embedded in the enforcement bureaucracy. This requires policies clear enough to secure compliance even in the face of grudging disagreement with those policies, but flexible enough to reflect the variety of situations that can come up in real life. Ideally, these policies would have an empirical basis as well.
For example, a policy might identify specific aspects of a case that have support in empirical evidence (e.g. history of prior calls, signs of physical injury, intoxiciation, identity of person calling for help, etc.) that warrant low, moderate or great suspicion regarding a domestic violence victim's capacity to act autonomously, and treat the situations differently, with great deferrence given to a possible victim's express statements in low suspicion cases, and something close to a zero tolerance/mandatory arrest regime used in high suspicion cases. In situations of moderate suspicion, procedural methods (like insisting on the involvement of attorneys in mediations and on private interviews with suspected victims in a safe place after the incident is over) might be used to discern a victim's real intentions. There is nothing sacred about that particular policy, which I simply made up while I was writing this, but it is illustrative of what a policy of that type might look like in practice.
* Per Wikipedia:
The dominance model and its most notable proponent, Catharine MacKinnon (b. 1946), reject liberal feminism and view the legal system as a mechanism for the perpetuation of what she sees as male dominance. It thus joins critical legal theory, which often underlines the instrumentalization of law for domination purposes.
Sexuality is central to MacKinnon’s dominance account. She argues that women’s sexuality is socially constructed by male dominance and that women’s subordination results primarily from the sexual dominance of women by men. A major line of critique of the dominance model is that it leaves no room for women’s agency since women are victims whose lives are fundamentally shaped by male dominance. The difficulty of the dominance model is that it leaves little or no room for women’s genuine autonomy or agency. . . .
Feminists from the postmodern camp have deconstructed the notions of objectivity and neutrality, claiming that every perspective is socially situated. Anti-essentialist and intersectionalist critiques of feminist jurisprudence have objected to the idea that there can be any universal women’s voice and have criticized feminists, as did Black feminism, for implicitly basing their work on the experiences of white, middle class, heterosexual women. The anti-essentialist and intersectionalist project has been to explore the ways in which race, class, sexual orientation, and other axes of subordination interplay with gender and to uncover the implicit, detrimental assumptions that have often been employed in feminist theory.
While my reasons for believing it is sometimes inappropriate to disregard the express preferrence statements made by women in relationships with domestic violence differ somewhat from the more theoretical model of dominance feminists, the phenomena of women incapable of protecting themselves in a relationship, even when they have brief opportunities to escape the relationship and fail to do so, is real enough.