Carceral and Intersectional Feminism in Congress

This paper uses a materialist feminist discourse analysis to examine how women’s movement organizations, liberal Democrats, and conservative Republican legislators shaped the Violence Against Women Act (VAWA) and the consequences for intersectional and carceral feminism. Drawing on qualitative analysis of Congressional hearings, published feminist and conservative discussion of VAWA, and accounts of feminist mobilization around VAWA, I first show how a multi-issue coalition led by feminists shaped VAWA. Second, I show how discourses of crime intermixed with feminism into a polysemic gendered crime frame that facilitated cross-ideological support. Third, I show how, in contrast, intersectional issues that activists understood as central to violence against women were discursively and structurally separated from gendered crime in Congress. Although a multi-issue movement coalition advocated for expansions in VAWA dealing with immigrants, unmarried partners, same-sex partners, transgender people, and Native Americans, these issues were understood in Congress through more controversial single-issue discourses and often considered in administratively separate Congressional committees. Fourth, I show how VAWA’s outcomes played out in terms of carceral and intersectional feminist goals.

AUTHOR'S NOTE: Many thanks to Carrie Baker, Steve Boutcher, Nicole Doerr, Marc Steinberg, Deputy Editor Krista Brumley, and

the anonymous reviewers for comments.
Correspondence to: Nancy Whittier, 10 Prospect St., Smith College, Northampton MA 01063. Feminists have held up the Violence Against Women Act (VAWA) as an example of both the promise and the perils of federal policy for improving the lives of women. Passed in 1994Passed in and reauthorized in 2000Passed in , 2006Passed in , and 2013, VAWA is the major federal legislation on rape and domestic violence. Co-sponsored by Senators Joseph Biden (D) and Orrin Hatch (R), VAWA was written by Biden's senatorial staff in close cooperation with the National Organization for Women's Legal Defense and Education Fund (Strebeigh 2008). Yet, prominent conservatives in Congress actively supported it. The collaboration between women's movement organizations, liberal Democrats, and conservative Republican legislators rested on a frame of gendered crime that invoked both feminist and criminal justice discourses and could be understood differently by different audiences. While this gendered crime frame for understanding violence against women was widely supported over time, intersectional feminist frames were more contested. A multi-issue coalition advocated for provisions in VAWA to address sexual and intimate partner violence against immigrants, same-sex partners, transgender people, and Native Americans. These issues fit into intersectional feminist discourse about gender and violence, but not into the gendered crime frame that facilitated conservative support for VAWA. The intersectional feminist frames that underlay activism for these expansions were discursively and administratively marginalized in Congress. They were debated through more contested single-issue discourses, often in administratively separate Congressional subcommittees.
A focus on gendered crime is consistent with what has been termed carceral feminism (Bernstein 2012). Carceral feminism refers to feminist activism aimed at increasing state enforcement against violence against women. What Bernstein (2010, 55-6), paraphrasing Bumiller (2008), calls the "neoliberal sexual violence agenda of feminism" entails a turn away from structural and economic transformation toward protecting women from a "menace [viewed as] squarely outside the home." Enlisting the punitive state rather than the welfare state, carceral feminism conceptualizes "social justice as criminal justice" (Bernstein 2010, 57). Carceral feminism is a term of critique meant to point out the dangers of relying on the state's punitive power to advance women's liberation. Few, if any, activists on any issue identify as carceral feminists. Bernstein (2012) has shown empirically how carceral feminism characterizes activism against prostitution and sex trafficking, but with little empirical study the term has been extended broadly to activism against violence against women. .
In contrast, an intersectional feminist approach emphasizes how social, economic, and political forces interact to shape different experiences and necessary solutions to violence (Arnold 2013;Naples 2009). For example, immigrant women are not only at greater risk of sexual and domestic violence; they are vulnerable in particular ways because of their precarious legal standing, language barriers, lack of access and understanding of the U.S. social and legal systems, dependence on husbands, and economic precarity (Bhuyan 2008;Ammar et al. 2005;Chen 2000;Villalon 2010). Law enforcement responses to violence against women can perpetuate violence against groups that are heavily policed based on race, class, or immigration status, while social services that treat violence primarily in terms of gender do not work well for women of color, non-English speakers, immigrants, and low income women (Richie 2012;Crenshaw 1991;Gillum 2009;Presser and Gaarder 2000). Some activists against violence against women therefore argue that state intervention, including funding, inherently undermines the interests of women of color or other marginalized groups (Bierria 2007). Others disagree and instead seek to alter the form of state intervention (Arnold 2013;Chen 2000;Villalon 2011).
In this study, I show that most feminists working for VAWA saw violence against women as a matter of gender inequality, not law and order. They did not identify as carceral feminists and many took an intersectional approach. Nevertheless, I argue that their activism around VAWA incorporated elements of both carceral and intersectional feminisms. They strategically used a frame of gendered crime alongside other frames.
They sought and achieved both social and criminal remedies and specialized responses for immigrants, Native Americans, women of color, and others; and they were grounded in a coalitional feminist movement. As a social movement outcome, however, VAWA, like most movement outcomes, fell short of what activists sought. Using materialist feminist discourse analysis (Naples 2002), I unpack how activists' goals and frames were refracted through existing discourses and structures in Congress that favored single-issue over intersectional models of gender, and criminal justice over other remedies, shaping both discourse about VAWA and the law itself.
The next section of the paper outlines the materialist feminist discourse analysis approach and is followed by an overview of methods. The substantive analysis, which follows methods, unfolds in four sections: First, I show how a multi-issue coalition led by feminists shaped VAWA. Second, I use materialist feminist discourse analysis to unpack how discourses of crime mixed with feminism into a gendered crime frame that facilitated cross-ideological support. Third, I show how, in contrast, intersectional issues that advocates understood as central to violence against women were discursively and structurally separatedand thus more controversialin Congress. Fourth, I show that VAWA's outcomes included carceral, non-carceral, and intersectional feminist elements.
Overall, I complicate the idea of carceral feminism by showing that feminist activism on VAWA combined carceral, non-carceral, and intersectional discourses, goals, and outcomes. Examining activism against VAWA as a social movement, I argue, shows that it reflects discursive and political constraints as much as activists' goals.

AND CONGRESSIONAL HEARINGS
Theoretically, I draw on materialist feminist discourse analysis to analyze institutional processes within Congress. Nancy Naples, building on Dorothy Smith, argues for a materialist feminist discourse analysis, which unpacks how competing frames and discourses intersect with existing power relations. Both discursive fields and institutional structures affect how activists frame issues and how frames are adopted and altered by those in power (Naples 2002;Naples 2013;Smith 1999). Similarly, Myra Marx Ferree (2003) and Holly McCammon (2007) analyze "discursive opportunity structures," including institutional alignments and prevailing discourses. Discursive fields shape how activists' claims resonate within existing understandings (Rochon 1998;Snow and Benford 1998;Skrentny 2006;Crossley 2002). Taking a materialist feminist discourse analysis approach to Congressional hearings entails looking at the multiple frames deployed and at committee structure.
Although extensive scholarship focuses on how political opportunities shape movements' access to lawmakers (Meyer and Minkoff 2004;Kriesi 2004;Tarrow 1994), less work has been done to examine how discursive processes play out within legislatures. Congressional committee and subcommittee hearings are a major location for the construction and circulation of discourse and a setting where advocates make their case as witnesses (Sabatier 1991;Andrews and Edwards 2004;King, Bentele, and Soule 2007;Holyoke 2009). Lawmakers' conclusions are shaped by both testimony and their pre-existing beliefs about the issue (Skrentny 2006;Campbell 2002;Jacobs and Sobieraj 2007;Burns 2005). Particularly in early hearings on an issue, dominant frames and final legislative form are not predetermined and testimony from advocacy groups can shape policy (Lohmann 1998;Burstein and Hirsh 2007;Banaszak 2010;Soule and King 2006).
Legislators and witnesses interactively construct the meaning and worthiness of potential policies (Allahyari 1997). Framing in hearings is strategic, aimed at persuading diverse audiences; groups that differ politically may use similar language for different ends (Naples 1997).
Congressional committee structure defines areas of jurisdiction that favor separate over intersectional issues (Davidson et al. 2014). For example, legislation pertaining to Native Americans is assigned to Indian Affairs, immigration to the House and Senate Subcommittees on Immigration under the Judiciary Committees, and sexual assault and domestic violence are usually addressed in the Judiciary Committees' Subcommittees on Crime. This institutional structure reinforces the use of separate frames of Native American sovereignty and immigration to understand violence against indigenous women and immigrants, and their separation from the single-issue gendered crime frame used for VAWA as a whole. By analyzing both the structure of committee jurisdiction and the frames used by witnesses and legislators, I show how VAWA unfolded along lines that emphasized gendered crime rather than intersectional feminism.

METHODS
The paper draws on qualitative analysis of Congressional hearings and accounts of feminist mobilization around VAWA. Data on frames in Congress consist of transcripts of Congressional committee and subcommittee hearings. These are a rich source of data on how witnesses, organizations, and lawmakers frame their positions (Brasher 2006). I compiled all committee hearings on VAWA from the first hearing in 1990 through 2012 through a subject search on the Proquest Congressional database. This yielded thirty-one hearings with 241 witnesses. All thirty-one hearings were coded for date, chamber (House or Senate), subcommittee, party control, and witnesses' affiliation, gender, and position on VAWA. I read all thirty-one hearings for background understanding and report here on their overall patterns. The analysis of frames draws from a subsample of thirteen hearings. To construct the sample, I selected nine of the nineteen hearings that are part of VAWA's legislative history, meaning that VAWA was assigned to that committee for official consideration. Specifically, I selected hearings from all four reauthorization cycles and from both House and Senate in cycles when both chambers held hearings. I selected the longest hearings from each cycle and purposively sampled all hearings at which witnesses affiliated with women's movement organizations appeared. I added four other hearings that addressed VAWA outside the official legislative history, three because they were especially lengthy, and one in the Senate Committee on Indian Affairs because of the topic. Hearings in the qualitative sample, for which full information is in the references, are: U.S. House 1992House , 1993aHouse , 1993bHouse , 1994House , 1999House , 2000and U.S. Senate 1990aand U.S. Senate , 1990band U.S. Senate , 1991and U.S. Senate , 1996and U.S. Senate , 2005and U.S. Senate , 2007and U.S. Senate , 2009 Using Atlas.ti, I coded statements by witnesses and legislators for frame. I generated frame codes inductively, then developed criteria for coding specific frames, which I did on a second pass through the data. I coded frames as crime, gendered crime, feminist, intersectional, and several specific frames like immigration. I did not code for carceral feminism because of the inductive approach and in order to capture variation in use of crime and feminist discourses. To construct these frames analytically, I used several criteria. Statements were coded as crime when they focused on law enforcement without inclusion of feminist claims. Crime frames discussed crimes against women as an instance of increasing violent crime in general, including the idea of an epidemic of violent crime, the need to crack down on crime, rising crime rates, and fear of juvenile offenders, all of which were part of the general crime discourse of the 1990s. Gendered crime frames included the importance of or need for law enforcement responses and/or statements reaffirming that violence against women is a crime, and they included at least some claims promulgated by feminists, including the widespread occurrence of violence against women in the family, effects on women such as fear or restriction of activities, gender norms, or male domination (Corrigan 2013;Weldon 2002). Simply mentioning women was not enough to classify a statement as using a gendered crime frame, but stating that "too many women live in fear in their own homes" was. To be classified as using a feminist frame, statements had to specifically reference gendered power or control or the effects of the gender system on perpetuating violence against women and not emphasize crime. Many speakers used more than one frame in a hearing, all of which I coded.
The gendered crime frame was broad, implicitly referencing different underlying ideas or discourses about gender, such as the chivalrous idea that women need better protection, the idea that women are dominated by men in a sexist society, or the need for less gender-biased law enforcement so that violence would not restrict women's freedom.
The underlying idea that women's social position shapes crimes against them linked these diverse ideas. The crime frame can be considered "carceral" and the gendered crime frame "carceral feminist." Both emphasize the need for state punishment and enforcement, but vary in their additional emphasis on gender. The feminist frame code does not emphasize carceral remedies.
I also coded for intersectional feminist frames, operationalized as explanations of how statuses besides gender affect experiences of or desired responses to violence against women. These were rare. I also coded for any mention of race, LGBT issues, immigration, Native Americans, or other specific groups, like youth or the elderly. There were almost no mentions of race or specific groups and none of LGBT issues. I therefore focused on immigration and Native Americans. I coded discussions of immigration and Native American women for intersectional or issue-specific frames of general immigration issues or Tribal sovereignty. I also coded for other frames, including public health and fiscal savings, which were rare. I looked for frames opposing VAWA, but none appeared in the hearings. No opposing witnesses testified except around immigration provisions, nor did legislators openly oppose VAWA. Most of the disagreement over VAWA's components occurred offstage, as legislators negotiated funding levels, specific provisions, and trade-offs for support (Strebeigh 2009).
I summarized the number of hearings in the qualitative sample in which given codes appeared at least once. This rough measure does not account for how many times a code appeared in a hearing or the length of coded statements. For example, all hearings contained comments coded "gendered crime," which appeared from 1 to 18 times; comments varied in length and some speakers made multiple comments (each coded as a separate instance). I also coded the frames that each speaker used in a given hearing.
Speakers could be coded as using one or more frames in a hearing, for example, as using only crime frames, or crime and gendered crime frames.
For data on the activities of feminist and allied activists on the issue, I draw from a 2009 symposium on VAWA that included the key players who drafted VAWA and worked on it over the years (cited here as Anonymous 2010) and secondary accounts. I use position papers produced by the major organization working for VAWA, the National Task Force to End Sexual and Domestic Violence Against Women (which I located through that organization's web page), to discuss organizers' frames and talking points. The law's centerpiece, in the view of Biden and feminist supporters, was Title III, which allowed women to sue attackers in federal court if they could show that the attack was motivated by gender. Title III was overturned in 2000 by the Supreme Court (Strebeigh 2009). Title III was symbolically important because it recognized violence against women as a matter of civil rights, but its scope and potential impactalthough significantwere relatively minor compared to the many elements and extensive funding contained in the rest of the Act.
Feminists working for VAWA were grounded in an intersectional coalition.
Biden's staffer Nourse, NOW's Goldfarb, and Helen Neuborne assembled a multiorganization task force headed by longtime women's movement organizer Pat Reuss to work for VAWA's passage. Members helped draft the legislation, organize Congressional testimony, and lobby for passage (Strebeigh 2009). After VAWA passed in 1994, the group formalized as The National Task Force to End Sexual and Domestic Violence Against Women and continued working to revise and promote the legislation.
Besides NOW, the Task Force included most major anti-violence and victims' rights groups; the major national and local "civil rights, labor, religious, youth, and community organizations;" and many immigrant, refugee, and anti-poverty groups (Anonymous 2010, 521, 528). Reuss said she sought to make the Task Force as diverse as possible, aiming "to make sure that every voice and every need and every group is at the table …women of color, disabled women, older women, women in the military…" (Anonymous 2010, 528). Lisalyn Jacobs, formerly a staff member in the Federal Office of Violence Against Women, chaired the group working on the 2006 VAWA reauthorization, which sought broader provisions for immigrants and Native Americans.
She explained that "we figured out what other communities needed to be brought to the table," including "the elder communities, the rural communities, the communities of color, the immigrant communities" (Anonymous 2010, 577-8). The Task Force eventually developed working groups focusing on different aspects of VAWA, such as immigration provisions, which were made up of groups that worked with the relevant issues or populations, such as immigrants' rights and immigrant women's organizations.
As Leslye Orloff, who had a long history in immigration activism put it, the coalition included "people…who hold us accountable and make sure that what we were crafting would really work for immigrant victims" as well as "incredibly strong mainstream allies" (Anonymous 2010, 582; National Task Force, n.d.). Despite the criminal justice elements of VAWA, this intersectional coalition cannot be considered simply carceral feminist, both because of its composition and because of the breadth of its goals.
Congressional committee chairs have authority over scheduling hearings and calling witnesses, but their staff make most of the arrangements and often rely on other groups for help (Davidson et al. 2014;Miller 2004). NOW and the Task Force worked with both Republican and Democratic committee chairs, suggesting language, providing supporting data, recruiting and prepping witnesses, and lobbying. In addition to hearings, Congressional staff and lawmakers gather information, draft legislation, and negotiate behind the scenes, in mark-up sessions, and in the full Congress after legislation is reported out of committee. The Task Force produced many short summaries of issues that provided talking points for hearing testimony and these backstage negotiations.
Despite bipartisan support, Republicans convened fewer hearings on VAWA than Democrats. Only 23 percent of all 31 VAWA hearings were held under Republican control, although Republicans controlled the House or Senate 52 percent of the time.
Republicans also called fewer witnesses per hearing on average (5.7) than Democrats (8.3). There was no difference in the gender of witnesses under Democrats (69 percent women) and Republicans (71 percent), and only small and unpatterned differences in witness affiliation. Despite their involvement behind the scenes, only six witnesses from NOW or the Task Force testified, at four hearings (all of which are in my qualitative sample), this was likely a strategic choice to make conservative support more palatable (Chen 2000).

FRAMING GENDERED CRIME
Gendered crime and crime were the major frames for understanding domestic violence and rape and were used by members of Congress or witnesses in all thirteen hearings sampled. Feminist frames, in contrast were used in only six hearings. Overall, 83 percent of speakers, including witnesses and legislators, made statements that were coded as crime, gendered crime, feminist, or a combination of these. 2 (The figures in the following discussion refer to the percentage of speakers whose statements in a given hearing drew on at least one of these frames.) Feminist frames were not used in the sampled hearings after 2000, that is, during the latter two reauthorizations. Gendered crime frames were used during all reauthorizations by between 70 and 22 percent of speakers, while crime frames were used by between 78 and 38 percent of speakers. 3 Change over time in the use of gendered crime and crime frames in the hearings is not linear and does not correspond to Republican vs. Democratic control of Congress. The much smaller number of witnesses in the latter two reauthorization periods (8 and 9 respectively) in comparison to the first two periods (83 and 30 respectively) also makes systematic comparison of frames problematic. I have thus combined hearings across time in the following analysis. Next, I discuss how witnesses and legislators used these frames in hearings and how their points of overlap fostered broad support for VAWA.
Democrats, Republicans, witnesses from women's movement organizations, and witnesses from sexual and domestic violence organizations differed in the frames they used.
Crime frames emphasized that domestic violence and rape are common, part of a "violent crime epidemic" (U.S. Senate 1990a,19), and equated violence against women to other crimes. Republican Senator Charles Grassley, a co-sponsor of the bill, exemplified this approach, decrying widespread "crimes of violence, whether committed against an elderly pensioner or a child abused by a drug-addicted parent, or against women, the subject of this hearing" (U.S. Senate 1991). Supporters called for greater law enforcement on the grounds that, "batterers must be treated like the criminals that they are" (U.S. Senate 1994, 3). Republicans were the most frequent users of crime frames.
The vast majority of Republicans used crime frames alone (55 percent) or in combination with other frames (75 percent), as did half of Democrats (33 percent used crime frames alone). Over half (56 percent) of witnesses from sexual or domestic violence organizations, such as shelters, used crime framing alone or in combination with other frames. For example, Beverly Dusso from the Harriet Tubman Center, a family violence service agency, used a crime frame to tout, "the really positive effects of the tough new penalties assuring families necessary Federal protection" (U.S. Senate 1996, 45). No witnesses from women's movement organizations used crime frames. I therefore do not classify crime frames as carceral feminist, because they are not articulated by feminist organizations.
Crime frames did not invoke a gender analysis. Whereas some conservative speakers may have eschewed a gender analysis in order to promote a broad crackdown on crime, some advocates did so in order to garner support for what could otherwise be dismissed as a matter of gender, not crime. As Maryland legislator Connie DeJuliis stated directly: "[T]his issue …is not a woman's issue… It is a crime issue" (U.S. House 1994, 21). As with other crime, the solution to violence against women in this frame was arrest, conviction, and long sentences. This was consistent with the crackdown on crime during the 1990s and 2000s.
In contrast, feminist frames were most common in testimony from women's movement organizations. All of the six such witnesses used feminist frames alone (33 percent) or combined with gendered crime (67 percent). For example, Kathryn Rodgers, Director of NOW-LDEF, stated that the group was "committed to protecting women's rights, and trying to eradicate violence against women has been at the top of our agenda since our inception 26 years ago." She described goals including, "creating a sea change in our cultural norms in this country" (U.S. Senate 1996, 75). Some service providers and survivors also used feminist frames, mainly when recounting personal narratives. Overall, lawmakers, like witnesses from anti-violence organizations, rarely used feminist frames. The emphasis on gender much more often emerged in a hybrid understanding of violence against women as a gendered crime.
The gendered crime frame was capacious enough to justify broad support for VAWA, and could be understood in terms of crime or feminism. Statements using a gendered crime frame varied in how much they emphasized crime (i.e., mandatory minimum sentencing, harsh punishment as just and deterrent, the threat of crime) and gender (i.e., the structural and cultural circumstances shaping violence against women and its effects on women as a group). Fifty-eight percent of all speakers employed a gendered crime frame in at least part of their testimony. Sixty-three percent of witnesses from anti-violence organizations used a gendered crime frame alone or in combination with other frames. Women who had been assaulted stressed how lack of enforcement and support services had affected them. For example, Yvette Benguerel recounted how law enforcement failed to intervene when her husband attempted to murder her and declared, "Women are not safe in this country, not even from the people they love" (U.S. House 1994, 14). Law enforcement witnesses, who often headed sex crimes or domestic violence units, also often used a gendered crime frame to emphasize the role of gender inequality, decrying the "culturally entrenched" attitudes that support domestic violence, as a prosecutor in a Baltimore domestic violence unit put it (U.S. Senate 1990a, 85). Two-thirds of witnesses from women's movement organizations used the gendered crime frame, often at the same hearing where they also used feminist frames. For example, NOW's Helen Neuborne decried that the "epidemic of violence against women is depriving half of America's citizens of their most basic civil rights" (U.S. Senate 1990, 57). (combine with paragraph below?) Similarly, 69 percent of Democratic legislators and 41 percent of Republicans used the gendered crime frame. Senator Biden's statement is typical, "[R]ape is a crime of hate, not of sexual desire. … battering is a crime of force, not of domestic discord…. These are crimes of terror. They instill fear not only in the actual survivors but in every woman in America" (U.S. Senate 1991, 1). Speakers using a gendered crime frame cast arrest as changing the cultural acceptability of men's control over women. As Republican Representative Steven Schiff explained, domestic violence in the past "arose from a feeling that men owned their wives and could take any action they wanted to as virtual slave owners. But that is long past. It is time that enforcement of the law caught up with that" (U.S. House 1994, 5).
The gendered crime frame was compatible with the other frames because, like the crime frame, it drew on crime discourse about the need for increased enforcement against violent crime and, like both crime and feminist frames, it sought to include violence against womenincluding in the homeas a violent crime. Even lawmakers and witnesses who did not use the gendered crime frame could and did agree with those who did. No one challenged any of these frames or their tenets in any of the hearings I coded.
The gendered crime frame can be considered carceral feminist because of its central incorporation of law enforcement as a remedy, its definition of violence against women as a crime, and its single-issue rather than intersectional approach to gender. However, activists used it to promote outcomes that were both carceral and non-carceral, as I will illustrate below.
Other frames circulated outside Congress, including what Berns (2001) calls the patriarchal-resistance perspective, which denies the prevalence of violence against women. Senator Biden often referred to the "hate mail" he received from men and "very fundamentalist churches" (U.S. Senate 1996, 50). These alternate frames never appeared in witness testimony or lawmakers' statements. Even Title III, which was not universally

BARRIERS TO INTERSECTIONAL FEMINISM
The VAWA coalition sought to expand and improve services like culturally appropriate shelters, and it advocated to better address the specific needs of women of color; immigrants; youth and the elderly; LGBT people; low income, rural, and disabled people; and others who are marginalized by government and mainstream services (Crenshaw 1991;Gillum 2009;Weldon 2002). Framing these changes in terms of "underserved populations," the Task Force wrote incentives into the law for organizations to develop specialized programming. Orloff reported that Biden's staffer Nourse: literally handed me the bill and said, "Tell us what we need to include to help immigrant, native women and women of color victims.' I …created the first draft of the underserved victim definition of VAWA. We …wrote access for underserved victims into the bill everywhere we could think of" (Anonymous 2010, 581).
While the term "underserved populations" could imply simply including excluded groups, advocates understood these provisions through an intersectional, not simply additive lens. For example, a Task Force fact sheet explained that underserved groups faced "unique challenges, reluctance to seek support, and few gateways to services," and that "victims often look for assistance from programs in their neighborhoods such as youth centers, senior centers, immigrant and cultural organizations … designed to specifically serve the community with which the victim identifies" (Lovelace 2012).
Congressional hearings, however, saw little intersectional feminist framing, even from advocates, and little discussion of underserved populations, who were addressed mainly through the Office of Violence Against Women's administrative and grant procedures (Chen 2000). Despite conservative opposition to a clause prohibiting discrimination in services against LGBT people, there was no discussion of LGBT issues in the hearings.
The only issues addressed at length were immigration and Native American issues. I have therefore chosen these issues as case studies of how intersectional frames and issues played out. As I show next, in hearings these issues often were framed as separate from violence against women, rather than intersectionally or as gendered crime, and they were administratively separated from the rest of VAWA through Congressional committee structure.

Immigration
Activists focusing on immigration emphasized how immigration status, ethnicity, language, class, and gender affected victims, and they sought measures protecting immigrant women from deportation if these women reported domestic violence or sexual abuse (Ammar et al. 2005;Bierria 2007). The inclusion of such measures came directly from the intersectional feminist coalition pushing for VAWA, which included the National Network to End Violence Against Immigrant Women and several other immigrant women's groups. According to Shana Chen's (2000) interviews with participants, in an attempt to bypass conservative opposition to immigration reform activists strategically decided to have anti-violence activists take a more public role than immigration activists in Congressional testimony and lobbying.. That is, advocates sought to frame the provisions in terms of gendered crime, not immigration. Very little discussion about immigration occurred in the hearings leading up to VAWA's initial passage, and activists did ultimately secure bipartisan support for limited visa provisions in the initial VAWA. As VAWA was revised, provisions for immigrant women expanded. Fairly narrow in scope, they nevertheless led to sizable numbers of visas granted with modest increases each year (Bhuyan 2008;Anonymous 2010, 583).
In contrast to the minimal differences in frames used by Republican and Democratic lawmakers and feminist witnesses in most VAWA hearings, differences were stark in the case of immigration. Intersectional frames were infrequent and were used only by witnesses from advocacy or service organizations and occasional Democratic legislators. For example, Beverly Dusso of the Harriet Tubman Center argued, "The horrors these families face are well beyond typical language and economic barriers. They have the added threats and retaliation and retribution that can be brought to bear through their deportation or losing their children" (U.S. Senate 1996, 46). Leslye Orloff explained succinctly, "[C]ontrol over immigration status, intimidation, isolation with language, cultural barriers, all accentuate the ability to abuse" (U.S. House 2000, 74). Supportive language was sometimes additive, as when a witness from a Florida shelter explained, "[I]mmigrant women have special issues. They suffer even more fears, more threats, intimidation and isolation than their American sisters" (U.S. House 2000, 65). The underlying analysis, however, was consistently intersectional, as witnesses emphasized how immigration status, gender, economic marginalization, and national origin together shaped distinct experiences of abuse (Chen 2000; National Task Force 2013; National Network on Behalf of Battered Immigrant Women, 2014).
In contrast, Republican lawmakers used only narrow immigration frames that did not focus on violence against women, raising objections to any path to legal status for undocumented immigrants and concern over whether VAWA would lead to loopholes

Protections for Native Americans
Similarly, issues affecting Native American women were an intersectional matter for advocates, but a separate issue for Congress, and particularly for opponents. Native American women experience high rates of domestic violence and sexual assault and minimal law enforcement or social service response. Tribal lands have an autonomous but under-funded police and court system, and are subject to federal rather than state criminal law (Deer 2006). At issue, especially in the 2012 reauthorization, are assaults committed on reservations by non-Indians. Tribal police and courts did not have the authority to arrest or prosecute non-Indian offenders, such as spouses of Native women.
Discussion of violence against Native American women occurred almost exclusively in a 2007 hearing in the Senate Committee on Indian Affairs. It focused on the high rate of violence and the jurisdictional issues that stymie prosecution of non-Native offenders.
There was universal denunciation of violence against Native American women, framed in terms of crime by both Republican and Democratic legislators and all the witnesses at the hearing. For example, a witness from Amnesty International explained how jurisdictional and funding inadequacies "delay and prolong the process of investigating and prosecuting crimes of sexual violence" (U.S. Senate 2007, 7). Two witnesses from Native American women's organizations also put forward an intersectional analysis, emphasizing indigenous cultural values and practices. For example, Tammy Young, the Director of the Alaska Native Women's Coalition, described her group's approach as based in "our customary and traditional ways," explaining that the priority for "all of our family members … to be provided services…. is what in some ways sets us apart from non-Native agencies" (U.S. Senate 2007, 25). The Task Force put forth a similar intersectional analysis in its fact sheets, which called for "preventative cultural practices," "holistic rehabilitation…for victims and violent offenders," and "Elders' panels and tribal drug courts to restore an Indian sense of justice and fairness" (National Task Force 2012).
There was little overt disagreement in the hearing, which did not lead to a vote. Dispute Conservatives used frames of opposition to immigration reform and Native American sovereignty as a way of justifying their opposition to VAWA while proclaiming ongoing concern for violence against women. They opposed the provisions on immigration and Tribal courts with claims about immigration fraud and constitutional limitations to Native American jurisdiction. For example, Republican Senator Charles Grassley (2012) proclaimed, "Our substitute [bill] contains language that will reduce fraud and abuse…We cannot allow people to misuse the VAWA self-petitioning process to obtain a green card." In 2012, the political climate around immigration was polarized, with intense partisan debate over immigration reform. Because immigration was framed as a separate issue, not as a force shaping violence against women, opposition to immigration reform easily carried over into opposition to VAWA. Even longtime conservative supporters voted against the bill in 2012. The Senate version, endorsed by the Task Force, passed with only 15 Republicans voting yes. The Republican-controlled House passed its bill, opposed by the Task Force, along party lines, without the disputed provisions around immigration and Native Americans.
Democrats and feminist supporters of VAWA highlighted gender during the 2012 elections, using Republicans' opposition to VAWA as evidence of their "war on women." Conservative Congresspeople were in a discursive bind: they argued that VAWA's provisions for marginalized groups were about separate issues, not gendered crime, but liberal lawmakers and feminists amplified the gender discourse and public opinion agreed. After the 2012 elections, VAWA, including the Native American provisions and some expansions regarding immigration, ultimately passed with much less Republican support than in previous years (Gray 2013;Keen 2013). 4

CARCERAL AND INTERSECTIONAL FEMINIST OUTCOMES
Overall, VAWA increased criminal justice responses to domestic and sexual violence, but also increased shelters' and crisis centers' funding and influence over the state, promoted incorporation of the needs of women of color and other marginalized groups into shelters and services, and shifted discourse about violence against women.
Numerous local and regional experts reported that VAWA increased federal and state prosecution and conviction for domestic violence and sexual assault. For example, a Michigan prosecutor boasted that, after VAWA-funded training, he "got a conviction in every rape case I tried for almost two years afterward" (quoted in Anonymous 2010, This can be considered a carceral feminist outcome. Critics of VAWA point to its inclusion in the notorious 1994 crime bill, which included harsh mandatory sentencing guidelines and numerous measures strengthening the prison system, as an indication that it is embedded in a carceral approach (Rosenberg 2010;Brooks 1997). More generally, collaboration with law enforcement, state funding, and the discursive incorporation of violence against women as crime into the mainstream can also be understood as strengthening the punitive state (Bumiller 2008). In this view, service organizations that work with law enforcement become part of the state's social control and punitive systems, which are more likely to target women of color, immigrants, and queer people than to protect them (Richie 2012;Bierria 2007;Spade 2011). Further, the first two iterations of VAWA also contained incentives for law enforcement to adopt mandatory arrest policies, under which assailants must be arrested if police are called to a domestic violence incident. These were widely criticized for reducing victims' agency and accelerating the criminalization of men of color, and were eliminated in the 2006 reauthorization (Presser and Gaarder 2000;Rivera 1995Rivera -1996. Even those who on principle do not oppose state intervention raise important doubts about limited access to law enforcement remedies for undocumented immigrants, members of excessively policed groups, and victims who themselves have criminal histories (Bhuyan 2008;Burnham 2001;Villalon 2010). VAWA clearly reflected the state's growing punitive approach, but its outcomes also went beyond the carceral.
Many activists reported on collaboration between law enforcement and community rape crisis centers and shelters, which was favored by VAWA funding. Testifying before Congress in 1999, a shelter director proclaimed that VAWA "has created a very, very different environment …where people have been forced…to begin to work together, people who have been antagonistic; for example, police and domestic violence programs" (U.S. House 1999, 101-102). A Colorado advocate, Claudia Bayliff, described how, "the local district attorney allows the local rape crisis team to sit in … when they make … charging decisions" (Anonymous 2010, 574). For them, collaboration meant that law enforcement would take their perspective seriously, making prosecutions more likely to go forward, making police, forensic examiners, and others more supportive, and integrating services into the process. At the same time, rape crisis centers and shelters are less powerful than police and prosecutors, leaving them with no guarantee of influence; someespecially those serving immigrants or women of colorpreferred to avoid contact with police (Arnold 2013;Corrigan 2013). Better integrated community organizations did not convert law enforcement to intersectional feminism, but it is not solely a carceral feminist outcome, either. These changes, and advocates' support for them, do not reject gendered crime discourse. Instead, they seek to influence law and practice to give survivors and feminist service organizations more power.
The underserved populations component of VAWA provided a strong incentive for rape crisis and domestic violence organizations to expand services for non-English speakers, immigrants, people of color, homeless people, disabled people, men, children in domestic violence households, and other marginalized groups. Many did so, changing the landscape of shelters that had been prominently criticized for failing to accommodate those groups (Bhuyan 2008;Crenshaw 1991;Gillum 2009). As Orloff recounted, VAWA funding led to "program after program in communities of color, serving immigrant women, serving deaf women, serving rural communities, serving people that we never dreamed would ever have a part of the pie…" (Anonymous 2010, 583). Steering state funding toward more culturally appropriate responses reflects a strategy aimed at reshaping state and non-state responses along more intersectional feminist lines (Chen 2000;Villalon 2011;Rivera 1995Rivera -1996. While advocates used gendered crime frames to gain votes, they also gained some intersectional outcomes. In terms of discursive outcomes, VAWA advanced the idea of violence against women as both crime and a form of gender oppression. Biden's staffer Nourse believed that VAWA "changed the terms of the debate. The idea that women have a right to be free from violence directed at them because of their gender, which was once a novel concept, is now a mainstream, commonplace idea…"(Anonymous 2010, 522-3). This discursive outcome persisted even after its concrete manifestation -Title IIIwas overturned.
These cultural changes were a longtime goal of feminist activists (Arnold 2013). They are both carceral and non-carceral, but not explicitly intersectional.

CONCLUSION
In sum, the feminist coalition that helped write and organize for the passage of VAWA was an intersectional one, in which race, citizenship, and other factors were understood as shaping how women experienced sexual and intimate partner violence, what they needed in response, and how state interventions fell short. But the legislation itself reflected an official discourse that understood violence against women uni-dimensionally in terms of crime, not intersectionally. Feminists gained the support of both liberals and conservatives in Congress by collaboratively constructing a frame for understanding violence against women as a gendered crime that was compatible with multiple ideological positions. The gendered crime frame became a dominant way of understanding the issue, linking crime frames to feminist ones, and remained dominant despite Republican votes against VAWA in 2012.
The construction of violence against women as crime is over-determined; discursively because of the broad appeal of the idea of gendered crime, legally because the law defines sexual and domestic assault as crimes, and institutionally because of the powerful criminal justice system. Nevertheless, the racial, gender, and national politics of VAWA and feminist activism against violence against women do not only reflect a carceral feminism that overlooked how race and citizenship shape violence. VAWA activists targeted the state for change. They agreed that the criminal justice system failed to serve marginalized groups and soughtwith decidedly mixed resultsboth to improve it and to strengthen community organizations' engagement with women of color and immigrants. Their use of the gendered crime and crime frames in Congress did not reflect a simple embrace of the punitive state, but a strategic attempt to gain sufficient votes.
Their attempts to expand VAWA and promote intersectional feminist frames in Congress were limited primarily by the material and discursive context of Congress, not their own goals (Whittier 2009).
Materialist feminist discourse analysis unpacks how discursive and administrative structures limited intersectional feminist influence on the legislation. The issues of immigration and Native American tribal procedures could not be accommodated within the gendered crime frame, which highlighted women's vulnerability but not the structural vulnerabilities of race and citizenship. These issues were often also organizationally separated, considered in subcommittees on Immigration or Indian Affairs. A materialist feminist discourse analysis also points to both policy and discourse as important outcomes. Materially, VAWA increased both criminal justice and social service responses to violence against women and shifted their nature, including more funding for specialized services for marginalized groups. Discursively, VAWA advanced both the notion of domestic and sexual assault as crime and the idea that they were a form of gender oppression, but it did not advance an intersectional feminist understanding of violence against women.
The gendered crime frame remains powerful and shapes state actions in several related areas. Sexual assault in the military and on college campuses have sparked government inquiries, activism, and administrative action. In these cases, the gendered crime frame prevails; the issues are framed as an affront to women's rights and as crimes that have not been taken seriously. The dominant response to campus sexual assault and sexual assault in the military --that they be taken seriously and prosecuted (or administratively investigated and punished through state-mandated campus policies) --is a demand that the state use its punitive powers on behalf of women. This carries risks for women of color, undocumented immigrants, and other marginalized groups. For feminists, these campaigns, like VAWA, have both liberatory and regressive potential.
The VAWA case holds useful lessons for understanding the potential and limits of institutional social change. Even though advocates of VAWA were based in a diverse coalition and understood violence against women intersectionally, understandings of violence against women as uni-dimensionally gendered prevailed over intersectional ones in Congress. Even when advocates emphasized and achieved cultural and organizational change, they also bolstered the criminal justice system. The advocates themselves were by no means carceral feminist agents of neoliberal social control. But the discursive and material outcomes of VAWA were not determined by activists; they emerged from a twodirectional process of discursive and material compromise and coalition in Congress, in which both feminists and conservatives gained in some areas and lost in others.