The Act of Listening to “Battered” Women: An Ethnographic Comparison of Police and Emergency Responses in São Paulo, Brazil

In the context of increasing prosecution by victims of domestic violence, this article compares professionals’ responses to specific cases in two different institutional settings in São Paulo, women’s police stations and hospital emergency services. The article focuses on the first encounter of victims with policewomen and health care popular legal advocates. We take “interpretive relativism” (Geertz 1983) as a fundamental framework for comparing, ethnographically, two contrasting cognitive methods for adjudicating truth to the events narrated by the victims: the skeletonization of facts, typical of police officials, and the schematization of social action, typical of human rights practitioners. We conclude that while policewomen are ambivalent regarding women’s capacity to decide for themselves how to use the legal resources offered to them, popular legal advocates, in contrast, seek to empower women by improving their capacity to make well-informed decisions.

Police stations and hospital emergency units are the places to which people turn in situations of extreme vulnerability. However, making use of these services may have different implications. From the victim's point of view, it entails recognizing the legitimacy of one or both institutions for conflict resolution, once other means have been attempted or dismissed as possible immediate responses. From the institution's perspective, offering these services means to give political status to police and health care, once domestic violence has been recognized as a public policy issue requiring special treatment.
We discuss the nuances implied in the immediate responses to female victims of violence made by two different kinds of female public officers: policewomen and popular legal advocates (promotoras legais populares or PLPs). We focus our analysis on the performance of these practitioners in two different institutional settings: three women's police stations (delegacias de defesa da mulher or DDMs), one located in downtown São Paulo and two in the urban periphery, and an emergency unit (Núcleo de Atendimento a Vítimas de Violência Doméstica or NAVVID), within a public hospital also just outside the city.
Women's police stations, known as DDMs, were founded in the mid-1980s in the context of feminist and postdictatorship democratization movements. The first of their kind, Brazilian DDMs influenced similar The Act of Listening 98 By taking interpretive relativism as a fundamental framework, 4 we hope to connect judicial, counseling, and feminist perspectives, as well the victim's standpoint, in meaningful ways that can help us understand situated processes of decision-making, rather than focusing exclusively on whether female professionals recognize themselves as feminist or not, a point frequently discussed by feminist critiques to evaluate the effectiveness of female-to-female assistance (cf. Bonetti 2000;Hautzinger 1997;Nelson 1996;Santos 2005).
We try to answer the following questions: What visions of the law do policewomen and popular legal advocates mobilize when they construct their responses to victims? In formulating such responses, how is evidence constructed in each specific setting? As we conclude, policewomen are ambivalent regarding women's capacity to decide for themselves how to use the legal resources offered to them. In contrast, popular legal advocates seek to empower women by improving their capacity to make well-informed decisions. Nonetheless, what could be a sensitive listening to the cases turns out to be a skeletonization of facts, once legal advocates often take DDMs as a normative model for their practice.
First, we map out multiple interpretations and uses of the Maria da Penha Law, exploring contrastive (feminist and nonfeminist) legal epistemologies of domestic violence. We then describe the ethnographic methods used in each observed site, acknowledging the possibilities and limitations of each research experience. We contrast selected response situations that are relevant to the specific ways that policewomen and popular legal advocates use the law. Finally, we reflect on the relevance of relativizing notions of truthfulness and realness.

The Maria da Penha Law: Factual, Symbolic, and Pragmatic Interpretations
Adopting an interpretive perspective, as we do, involves approaching violence not as a thing but as a phenomenon. Even though Brazilian law provides a definition of family and domestic violence, we should be concerned not with fixing a definitive concept of domestic violence, but with comparing the different ways that legal/police officials and human rights practitioners think of the concept. Indeed, the tendency to associate the term "victim of violence" with vulnerability, passivity, lack of autonomy, and inability to resist has been intensely debated, with some feminists and researchers suggesting the use of more "neutral" terms such as "women in situation of violence" (Lins 2014).
In his classic essay "Local Knowledge: Fact and Law in Comparative Perspective," Geertz (1983) refers to factualism as a contemporary Western legal phenomenon that has as its main consequence the polarization of facts and laws. Factualism is a paradoxical phenomenon: the more a fact fits the law, the more credible it is to the legal practitioner, yet the more distant it gets from what really happened. Geertz then develops a semiotic and comparative analysis of three legal traditions: Islamic law and its central concept of haqq, which roughly means "truth"; Indic law and its concept of dharma, which means "duty"; and Malaysian legal tradition and its concept of adat, which means "practice." In comparing those legal traditions, Geertz addresses the problem of how particular legal traditions adjudicate facts to laws, in an attempt to develop a cultural analysis that connects the intrinsic meanings contained in juridical codes with the particular contexts that give sense to them. By contrasting extremely different legal sensibilities, Geertz confronts how people in particular contexts get access, experience, and construct the law. It is in that sense that we would like to understand better what cognitive, moral, and legal frameworks turn out to be legitimate and valid in each institutional context, and how practitioners act upon those cases they configure as being fairly violent.
The Maria da Penha Law qualifies "domestic and family violence against women" as a heinous crime, entailing harsh punishment for the perpetrators and stronger protective measures for the victims. This penalizing trend stands in contrast to previous legislation (9.099/1995) that upheld conciliatory ideals, through which judicial officials decided in favor of reconciling the parties through faster conflict resolution procedures, either by persuading the victim to retract the complaint or sentencing the cases as minor offenses. 5 In line with the scale of the Brazilian Penal Code, jurisprudence generally involved alternative penalties, such as food basket donations or doing community services. As Debert and Oliveira (2007, 308-309) point out, at the Special Criminal Courts, proceedings were oriented toward the defense of the family, "reproducing the hierarchies and conflicts of that institution." To remedy the situation, in February 2012, the Brazilian Supreme Court sanctioned bodily harm complaints to be legally prosecuted regardless of the victim's wish. The decision, however, has produced docket overloads, more than doubling the number of investigations, which skyrocketed from 554 investigations in 2011 to 1,126 in the following year. Conversely, the number of police officers, the technical systems for filing complaints, and the clerks' wages have remained the same. Verbal assault, on the other hand, a minor offense yet a criminal category contained in the Brazilian Penal Code, requires the woman's formal proceeding along with a lawyer's appeal action for legal prosecution. As legal anthropologist Luiz Cardoso de Oliveira (2008) pointed out, Brazilian positivist law has great difficulty in understanding the dimensions of moral harm and the indignation and resentment felt by victims of abuse. To make matters worse, quite often the victims do not comply with the prosecution's demands, which entail long bureaucratic procedures such as providing trials or accounts by victims, perpetrators, and witnesses, and possibly medical evaluations, among other things, creating more obstacles for final sentencing (Lins 2014).
Alternatively, many scholars of the Maria da Penha Law agree, penalization apart, that the very law exerts an enormous "symbolic efficacy" (Pasinato 2008) on victims and legal practitioners. Maria da Penha Maia Fernandes, a Brazilian pharmacist whose husband's violence caused her to become paraplegic and who fought for him to be condemned, has become a powerful symbol, the very incarnation of domestic violence against women. 6 Today the law is one of the best known among the Brazilian population. 7 The name Maria da Penha is even confused by perpetrators with the places where victims go to register their complaints: "If you go to Maria da Penha, I will kill you," a male partner threatened his wife, referring to the police station where she said she would go to register his offenses.
In contrast to policewomen's factualism, the Maria da Penha Law's typology of forms of violence-that is, physical, sexual, psychological, moral, and patrimonial 8 -has proved extremely valuable to counseling practitioners, such as popular legal advocates. Capitalizing on the legal text to teach victims about the meanings of domestic violence has been essential for motivating women's confession of the truth by helping them understand their specific situations from a comprehensive legal perspective. As NAVVID's coordinator explained in a taped interview, "Our great effort is to show them [the victims] that those forms of violence exist." On the other hand, placing too many expectations on the power of the law as the only solution to women's problems can result in pitfalls, such as the tendency to configure a mode of assistance that relies strongly on information giving.
One of the main questions that emerges out of the international discussion on legal capacity concerns what epistemologies should orient popular legal practice, whether an alternative or feminist perspective. For instance, Ratna Kapur (1992, 95), who reported on an experience with a popular legal advocates' course in India, argues that legal capacity should be defined as making "women conscious about the limits and possibilities of the law." Kapur criticizes an "information-giving" standard model-very common in advocates' courses internationally-that relies strongly on the transference of information about legislation by means of flyers, seminars, short-term courses, or informal conversations, where women are approached as passive listeners.
It is worth noting that while popular legal advocates pursue the transference of knowledge about the law as an effective response to violence; police clerks, by systematically skeletonizing women's narratives for the sake of crime reporting, leave victims significantly uninformed.
Another important question originates in the international discussion on the feminist perspectives that should orient women's legal protection and defense. Guita Debert (2010, 479-481) highlights four main feminist perspectives vis-à-vis legal uses: liberal, which defends the equality of treatment between women and men; culturalist, which considers legitimate the asymmetrical application of equality; radical, which seeks to combat male chauvinism juridically; and postmodern, which highlights the diversity of the feminine experience, whether social, economic, racial, ethnic, religious, or age-based.
The Maria da Penha Law proves to be culturalist, at least from the juridical standpoint of "violence against women," a term coined by the feminist movement in the 1960s that presupposes violence as an effect of male domination. Yet, postmodern feminists, as well as policewomen, question victimization or the consideration of women "as passive victims of domination" (Debert and Gregori 2008, 167-168). Furthermore, even though the Maria da Penha Law appears to be radical, it does not seem so when viewed from the perspective of what feminist legal activists have pursued in other countries.
For instance, in Peru, where, like in Brazil, "being well treated by a public institution is still considered a favor rather than a right" (Dasso 1992, 179), paralegal courses are based on notions of alternative law and legal pluralism. Perú Mujer, a nonprofit organization located in northern Lima, adopts a pragmatic perspective to legal capacitation that includes teaching poor women skills such as how to distinguish substantive law from processual law; how to self-manage administrative judiciary paperwork, register complaints, activate, and follow up judicial processes; how to persuade police authorities on behalf of the women assisted; and how to gain public recognition by local governors (Dasso and Napurí 1991, 87).

Multisite Ethnographic Comparison: Methodological Considerations
To do an ethnographic study of a single policy implementation process requires approaching multiple institutional sites. However, each case study began as a single case. We ended up multiplying our sites by aggregating other fields, not with the purpose of generalizing across cases, as a political scientist would do (Yanow, Schwartz-Shea, and Freitas 2010), but to learn something new by looking at one case through the lens of the other cases.
This comparative effort, as Geertz (1968Geertz ( , 1983 argued, is eminently hermeneutic and contextual and has multiple potentials, such as to observe mutual influences, to understand one reality more deeply through the other, to analyze apparent differences and elusive similarities, to develop an internal critique through an external analysis, or to evaluate one case through external cultural categories, among other plausible efforts. Because Geertz's approach to cultural systems is semiotic, and therefore actor-oriented and interpretive, institutions must be approached as the contexts from which meaning derives, and not as isolated entities that can be put in opposition to one other. Finally, even though ethnography is conventionally defined as an array of methods such as establishing rapport, observing situations, selecting informants, and transcribing texts, among other techniques, from an interpretive perspective, doing ethnography means thick description. As opposed to thin description, or the observation of an event as event, crafting thick descriptions of events is the interpretation of unobservable "meaningful structures" that nevertheless become intelligible through the observation of human action (Geertz 1973, 20-29). The point is trying to penetrate the way others think, or think they think. Be it knowledge, systems of thought, or legal traditions, for Geertz (1973, 5) one should, in the first instance, "look at what the practitioners of it do." Beatriz Lins conducted fieldwork between 2012 and 2013 in two of the DDMs studied, one located in the periphery of São Paulo and the other in the impoverished downtown area. Today, there are nine women's specialized police stations in the city. The research goal was to achieve a deeper understanding of the meanings and nuances of everyday police practices, particularly after the changes introduced by the Maria da Penha Law, by looking at the law as a dynamic and polysemous experience, whose meanings and uses may change and transform through the practices of different professionals (Lins 2014). DDMs were then open to the public only from Monday to Sunday, from 9 a.m. to 6 p.m. The field research included weekly visits in each DDM during a period of six months, and six tape-recorded in-depth interviews with the on-duty police clerks at each DDM setting.
Rocío Lorenzo carried out research in a hospital emergency ward between 2012 and 2016. Between November 2015 and January 2016, she expanded the field research to the DDM to which the majority of women assisted at NAVVID were referred. The on-site inquiry included direct observation of police clerks and victims' interlocutions, twice a week. Both the hospital and the DDM are located in a poor neighborhood on the eastern outskirts of São Paulo. The neighborhood is one of the five municipal subdistricts that make up the eastern zone of the city. With a population of 3,620,494 inhabitants, this area is the most densely populated region of São Paulo. The hospital assists approximately six hundred patients per day, according to hospital data.
In Brazil, the PLP project, as it is known, began in 1992, resulting in over two thousand women trained as popular legal advocates in São Paulo by 2015. The curriculum was based on the themes of law, human rights, women's rights, health, public security, gender and racial discrimination, domestic and family violence, education, and sexual rights. The women who attended the course were mostly lower class or poor and from fifteen to sixty-five years old; approximately half of them were white and half racially mixed. During a period of two years, Lorenzo observed the course activities, interviewed (informally) the women attending the course, and made regular visits to NAVVID, including tape-recorded semistructured interviews with the coordinator and the on-duty legal advocates who run that emergency unit.
Six women, among the hospital's public servants, were trained as popular legal advocates before they became NAVVID's on-duty attendants. They were then released from their previous tasks, which included work as receptionist, nursing aide, or laundry clerk. Their present work follows a rotation system that functions during the whole week, including the weekends, from 8 a.m. to 7 p.m., without having managed, so far, to keep the unit functioning at night. According to NAVVID sources, between October 2003 and October 2016, 2,657 cases of domestic violence were registered, reaching a peak in 2007 with 332 cases. Most of the victims were white and between twenty-one and thirty-nine years old. The causes of abuse, according to the victims' declarations, were drug or alcohol use by the perpetrators, betrayal, or jealousy. Sexual abuse, curiously, was the least frequent form of violence mentioned by victims.
The field methods used by both researchers are related to the different rules and procedures dominant in each institutional environment. Confidentiality is an important prerogative of medical assistance. As NAVVID's coordinator explained, "The popular legal advocates' job is extremely confidential because the victim can be abused a second time; she exposes herself too much by coming here. To betray confidentiality is a violence." This confidentiality diverges strikingly with the deliberate public approach of police officers to victims. From the first moment, when a woman finds herself at the reception desk, her speech is interpellated in public: doors are kept open most of the times and police officers circulate constantly among the rooms.
The two researchers, therefore, could not negotiate their on-site roles on equal "footing"-the alignment among participants in communicative interactions (Goffman 1979). While authorization to observe policewomen's conversations with victims was relatively easy to obtain, Lorenzo was not authorized by the hospital director to observe interactions between legal advocates and victims inside NAVVID. However, she was authorized to visit NAVVID at will, interview the on-duty attendants, and hang around and talk to people in the emergency facilities. Lorenzo also carried out thirteen in-depth interviews with victims, which were tape-recorded and transcribed.
In trying to contrast the different roles played by the two researchers in each institution, Goffman's (1979) distinction between dominating and subordinated communication turns out to be relevant. Goffman (1979, 7) distinguished different modes of being socially listened to, such as eavesdropping, pretending one is not listening; overhearing, hearing by chance; byplay, among official participants; crossplay, among official participants and spectators crossing the limits of dominant interaction; and sideplay, a whispering among spectators.
Despite not enjoying the official status of listeners, both of us were listening all the time, either eavesdropping or overhearing, mostly in waiting rooms, corridors, stairwells, and rooms where assistance took place. Byplay among clerks and other police officers was frequent, as among legal advocates and receptionists, nurses, social workers, and clients. Likewise, it was important to observe, mainly at the DDMs, how some victims initiated a sideplay with the researcher, who was observing assistance from a spectator's perspective, to whisper what was repressed by the clerk's dominant communication.
In sum, field research in each institutional setting was significantly marked by those constant and subtle interactions that, in the last instance, affected the kind of information collected, the methods used, and the interpretations produced in each study.

Constructing a Response
We pay special attention to the extent that decisions involving women who are assisted and those assisting them can be associated or disassociated in some meaningful way with the two cognitive methods proposed by Geertz (1983)-the skeletonization of facts and the schematization of social action. We distinguish between replies, that is, immediate face-to-face answers or reactions, and responses, that is, abstract and anticipated institutional responses, those that are not dependent on immediate face-to-face answers (Goffman 1981). Following interpretive relativism, we observe how parochial ways of exposing truth and realness by the victims and the professionals may emerge altogether from multiple local sources, such as social norms, moral expectations, intentionality, subjective experience, and abstract principles, as well as legal and feminist perspectives. We distinguish three kinds of responses to situations: (a) skeletonization of facts; (b) schematization of social action, and (c) referral to other institutions.

Skeletonization of Facts
In the following excerpt from a tape-recorded conversation between a female police clerk and a victim, a young racially mixed or brown (parda) woman is assisted in the clerks' room, in a DDM in the periphery of the city, on Monday, November 16, at 1:44 p.m. The clerk barges into the room and begins the conversation with a direct request: "Do you have an RG (identification card)?" As Ostermann (2002) concludes, by analyzing the "closing ups" of interactions between officers and victims at DDMs, assistance in this kind of service is oriented towards the production of an official document, the BO (boletim de ocorrência or police report), where the complaint is registered. Therefore, the organization of the conversation (its sequence) is characterized by the absence of classic polite phrases such as "goodbye" or "thank you," or, as would be expected in the following conversation opening, "hello," "how are you feeling today?" or "good afternoon": to kill me, and that he is going to disappear with my son. C: His son? V: Yes, we have a five-year-old son. He said that he is going to take him and that he is going to kill me and that if I come to the police, then I was going to see what's going to happen. C: Sunday now? Was it the eighth? V: It was the eighth. C: When did he say that he was going to kill you? V: It was the entire day. C: By phone? V: By phone. He went to my mother's house; he was eager to talk to her. C: But did he say to your mother that he was going to kill you? V: No, he told my mum that I didn't need to do that to him. C: But then, how is it that he said to you that he was going to kill you? V: That was to me. C: Was that personally or by phone? V: In person, he has already said this so many times, but that day it was by phone. C: Where were you? V: I work at a friend's shop. C: Where? V: Close to home. I sell Avon, Natura, that kind of thing. C: What time did he call you? V: It was around 10 or 11 a.m. C: Were you at the shop or at your mother's house? V: I was at the shop. She resells Natura, Avon products. C: On what street is the shop located? V: I live in Santo Antonio [fictitious name]. C: Did you receive the call at work or at home? V: So, he called me at 9:44 [she checks the cell phone]. He also called at 1:36, but I did not answer. C: But where were you when he called? V: I was going down the street where the shop is located. C: Were you on the street? V: I was on the street. C: 9:44? V: 9:44. C: Did he threaten you? V: He did threaten me. He said that he could not believe what I had done to him and that… [the clerk interrupts]. C: Were you on the street? V: I was on the street. C: What number, more or less? V: I was, more or less, passing through the market. C: Tell me an exact point. V: I was on the street. But he went to my mother's house [the clerk interrupts]. C: But didn't he threaten you on the phone? V: Yes, he called and threatened me. C: Is this street within an occupation [illegally occupied territory]? V: No, it is Santo X. C: I can't find it on the system. Is it a street or an avenue?
From then on, the conversation shifts from short questions and answers to the clerk simultaneous typing and reading of the police report. After thirty-eight minutes, the clerk reads the final report to the victim, who accepts it as valid. The criminal fact registered is "life threat," expressed in the following terms: "I am going to put an end to you, I am going to kill you. Is that right?" "Yeah." "And if you go to the police, you are going to see what happens. Right?" "Right." The victim informs the clerk that because she is frightened, she is moving to another state to live with a relative until the baby is born and that she does not want her ex-husband to be informed of her new address because he might accuse her of taking her son with her. The clerk replies that the Family Justice Court will decide what is best for the child. The victim signs eight forms and leaves the DDM.
In general, clerks begin by asking, "Ma'am, what happened?" and hear long, disconnected, and emotional outpourings about complicated love and family relationships. The police reports do not record a sequence of events, but only the last incident, a criminal fact, although they also register information about the relationship of the parties involved. Quite often, women have difficulty identifying only one incident or remembering the specific circumstances involved in the last time they were subject to violence. It is up to the clerk, therefore, to select what would correspond to an offense. Types of crimes, though, are unknown by most of the Brazilian population. Moreover, the decision regarding what is qualified as a crime can vary from one professional to another, particularly when the woman is unable to describe specific situations of beating, verbal assaults, or life threats.
The filing of a police report is an overloaded bureaucratic procedure that involves, above all, the computer and the frequently crashing system. The Complaint Police Report is the software used by the Civil Police for recording and classifying the data concerning offenses. In the preceding conversation, it took nine minutes for the clerk to find the name of the street where the victim said she had received the threatening call from her ex-husband. For this kind of attention that took exactly forty-six minutes, at a moment when there were eleven people in the waiting room (some of them accompanying the victims) and only two clerks on duty, nine minutes is a long time just for interacting with the system. As Michael Lipsky (1980, 30) claimed, for the case of street-level bureaucracies, "an emphasis on housekeeping chores, such as filling out forms … affects the amount of time available to clients." In the following excerpt from a tape-recorded interview, a sixty-year-old woman who suffered conjugal violence for twenty years tells the interviewer the reasons that led her to go to a DDM for the first time in her life to file a police report against her husband: So, he used to offend me and attack my daughter…. One night I told him, "Please, she is very nervous and mad at you, so, please, do not say anything about us in front of her…. And please, take that gun from that shelf [pointing to a shelf in the kitchen], because I am afraid she may do something with it." And he said, "If she does so, it is your fault!" That was the first step for me to register a BO, because he was jeopardizing my daughter. I thought, "My daughter taking a gun, killing herself, or shooting him …" I got frightened. So, that was on the weekend. On Monday I said, "Aren't you going to take that gun? Is it going to stay there?" I was really frightened. I went to register a BO against him at the woman's police station…. I said, "I don't know how I had the courage to come here! I don't know what will happen." Then, she [the clerk] said, "Do you want to prosecute?" And so on. I said, "No, no, no, I just had the courage to come here, I just want to register a police report to show him that I had the courage to do so." … When he found out, he didn't do anything. He only said, "So, you mean that you went to register a police report against me, did you?" "Of course," I said, "What were you thinking? I am not going to jeopardize my daughter because of your silliness, because you leave that gun there." … From then on, he took the gun from the shelf … Why was it necessary to do that? … Rocío, my daughter had said to me, … "I will not kill my father because I am already eighteen and I can go to jail." For God sake! … That is what led me to register a police report.
Next, we contrast the incident narrated by the victim in the interview to the incident registered by the clerk at the DDM Police Report, which she shared with the interviewer: The accused is a dishonest person, pretending that he is a nice person in front of others but is an aggressive person with the victim. She reports that she was assaulted on two occasions; however, she did not register the incidents. In the date of the facts, the accused locked himself in with the victim in a bedroom and began to question her about why she had talked to other people about their relationship. He also mentioned twice that he was going to give her a slap on her face, but he did not do it and uttered the following words: "That is why husbands kill wives like you. You spend your time babbling with others about our life, gossiper." She recalls that after approximately thirty minutes the accused left the room, without preventing the victim from getting out. The victim felt frightened and concerned with her physical safety and decided to register the incident.
Informed about the protective measures to which she has the right, she does not wish these to be applied in this act.
Being made aware of the 180-day term she has to prosecute, she states that she is certain that she does not wish to do so.
The victim went to the DDM with the single objective of registering a police report to make her husband take the gun from the kitchen shelf; she actually achieved this outcome very quickly, thanks to the BO. To do so, however, she skeletonized the facts in a way that guaranteed that her husband would not be incarcerated, not mentioning to the clerk anything about the gun.
In both cases, the victims showed great decision power in constructing a story believable or real enough that would allow the production of an effective police report, from their point of view. However, from the perspective of police efficiency, both cases represent a waste of time and resources. Police clerks are pressured by higher police officers who measure DDMs' efficiency according to the number of cases that result in concluding investigations. As the police clerks interviewed by Ostermann (2002, 43) confessed, they were often disappointed by those women who wanted to register a police report "just to have a document to take home," a kind of event they called "the yellow paper syndrome." Susana Durão (2013, 284) has suggested in her ethnographic study of Portuguese police squads that quite often, intersubjective exchanges lead police patrollers to adopt a "grammar of impotence" that results from their appropriation of the visions that others have of their work, which ends up dominating police patrollers and victims' moral expectations.
Many clerks, in fact, believe that one of the main problems of the DDMs is that women usually have different expectations with respect to what police work is about. As one clerk interviewed by Lins recognized, "some women just want to talk, they don't have the intention of actually prosecuting the man, they just want to unburden themselves…. Sometimes I let them do it, but that is not my job!" Likewise, a police investigator lamented, "We know that many people say that we are cold and treat women as if they were lying, that we do not give them humanized assistance. But the thing is, we are police officers, not psychologists! We are only interested in facts, not in whining."

Schematization of Social Action
NAVVID was created as part of a Humanization Project, an initiative carried out by the hospital director in 2003, with the goal of stimulating the "speaking and listening capacity" of health care professionals. 9 Observation, nonetheless, is the first ability that popular legal advocates must develop. As NAVVID's coordinator explained, "Popular legal advocates are trained to observe, not to ask." In a training meeting at the hospital, one of the six popular legal advocates who assist victims of domestic violence at NAVVID explained to a group of newly recruited receptionists how they must proceed regarding female victims of violence who arrive at the emergency ward: "NAVVID is a unit that assists women suffering from violence…. It is a difficult task: some arrive crying, screaming, others arrive mute. A woman who has suffered an assault never raises her head. When they have bruises, they wear sunglasses, and if they are accompanied by men who answer all the questions addressed to the women, these are clearly cases of domestic violence. When a woman arrives at the emergency ward, your role is to refer her to us. That is not what happens today: she opens a file and is referred to the surgeon, so she has already lost contact with us." The key to the legal advocates' job has to do with circumventing the emergency ward protocol. To do so, they need the consent and support of the hospital director, the doctors, the social workers, the nurses, and especially the security guards and receptionists, who, as gate keepers, are the ones who control who is treated and who is referred to other public health services. Yet the risk of losing clients in the physicians' hands is constant. Cynthia Sarti (2005, 114), while analyzing a unit for assisting victims of violence in a São Paulo public hospital, found that even when the physicians-who focus on the disease and not on the sick person-heard about the unit, "they did not recognize it as a legitimate service." Therefore, the job of popular legal advocate is extremely important at the reception desk, where they may identify potential victims. It is at the reception area that they circulate information about the Maria da Penha Law, by handing out an illustrated purple booklet that is easily comprehensible to the women who walk in or are waiting to be treated. The booklet contains the law in its entirety. The women who end up being treated at NAVVID are advised to carry the booklet with them always in case they have to go to a police station. As a legal advocate once vividly related, "I tell them, 'Scream, cry, shout, curse, do all you need to do here because if you go to a police station like that they are not going to take you seriously.'" At the emergency reception desk, violence can be both overtly visible and invisible. On the one hand, nowhere does violence acquire the ostensive visibility that it does in emergency wards (Deslandes 1999). On the other hand, it is quite common that victims arriving at the reception desk seriously injured lie to the person responsible for screening patients, saying that they have suffered an accident. Legal advocates registered one case involving a woman who arrived at the emergency ward with very serious burns and told the doctor who attended her that the kitchen gas container had exploded. With time, the legal advocate on duty was informed by the victim's daughter that she had been burnt by her husband. This woman died after twenty-six days of hospitalization.
In another observed situation, a legal advocate entered the NAVIDD's office distressed: "A case of violence has arrived!" A thirty-five-year-old woman had been stabbed by her thirty-nine-year-old partner, a crack user. The punctured arm was bleeding when she got to the reception desk, from which she was sent directly to the surgeon. When the advocate was informed about the case by a hospital nurse, the woman had already been checked into a room. The advocate went upstairs to talk to her and afterwards retold the story: "The girl said that her boyfriend went to the shower to wash the knife right after stabbing her arm and while he was doing that he kept kicking her. The social service worker has to call the police agent, who must come to the hospital to 'take testimony.'" On that occasion, both eavesdropping and overhearing were crucial for the researcher to understand what was going on.
At NAVVID, the rule is to "listen without making judgments," believing that the woman who arrives at the unit only wants to speak and that it is by means of her conversation that she will decide what action to take. As one of the legal advocates interviewed used to preach, "We try to listen first, look into her eyes, and put ourselves in her shoes…. We ask questions for them to reflect on…. How does he treat you? For how long has he been like that? Do you think he is going to change? Look at yourself: do you think you deserve that? Do you want your son to be an aggressor? Why don't you try a community therapy group?" Another legal advocate interviewed professed, "Women leave here empowered; our assistance is special, we listen, look at the woman with respect. The visit has no time limit; a professional was not going to be able to do that." One of the most involved advocates, who suffered conjugal violence herself, put it this way: "Their eyes [the victims'] brighten up when I look them in the eyes because they are not used to being treated with respect in a place like this." In practice, in the emergency ward, empathetic listening seems to be more closely associated with an ideal of mutual understanding with the goal of trying to put oneself into the other's shoes-as in Dilthey's classic empathetic transposition-than with a fully accomplished assistance procedure. As in the counseling interactions observed by Ostermann (2003Ostermann ( , 2002, the service offered in these kinds of first encounters is oriented toward the establishment of a professional relationship with the client. Therefore, politeness and affectionate ways of giving continuity to the conversation are encouraged. However, the emergency does not offer the conditions required for establishing continuity in client relationships.

Referral to Other Institutions
Despite valuing empathy, legal advocates seem also to be highly constrained by disrupting bureaucratic procedures, for they end up acting as if their ultimate goals were complaint filing, data gathering, and referral to other institutions, not very different from what the police do. Bureaucratization has been gradual, though. As a legal advocate revealed, "With the first woman I attended, I even accompanied her to the bus station." The first document they fill out is a handwritten short case description report, which is recorded in a paperboard cover notebook. A reproduction of this document follows: Case description: The client relates that she was assaulted by her partner. He beat her with a stove grill on the head. The aggression happened on the eighteenth of March and she only came to the doctor now because she could not stand the pain anymore. She said that has been living with him in matrimony for thirteen years and that she has eight children from this relationship. She says that he does not allow her to have friends nor visit her family. He offends and humiliates her in front of friends and neighbors. She said that she could not stand living this way any longer.
Advice given, police report registered, she is going to decide whether she goes to the [closest] DDM, referred to Maria House [closest women's help center]. Observation: the client continues under medical observation.
This case description report was handwritten by a legal advocate right after returning from the room where the physician was attending the victim. It differs from other documents in its presentation of a more humane style and because it does not have a predetermined bureaucratic endpoint, beyond its archival status into the hospital records. The complaint's open narrative format would presumably allow for a thicker description of the event, free of the system constraints that clerks face. However, this case description seems to be motivated by the single intention of registering a criminal fact.
It is exactly between what Geertz (1973) called "thin description," what the legal advocate describes as having happened to the victim ("He beat her with a stove grill on the head"), and the "thick description" of what happened ("she was assaulted by her partner") where there lies a stratified hierarchy of meaningful structures whose intelligibility-never complete, however-becomes more coherent the more contexts we aggregate in the story.
Still, popular legal advocates may end up filling up four more forms: (a) the "Report/Case Description," referred to the closest DDM, whenever the victim follows this recommendation; (b) the "Health Service Coordination Incident Report," from the São Paulo State Health Department, which the victim must hand in at a police station, with a copy of a medical care certificate, in case she decides to file a police Incident Report; (d) the "Violence Notification Record (Suspected or Confirmed)," to be recorded into the hospital archives; and (d) the "Suspected or Confirmed Case Notification," referred to the Violence and Accident Watch Information System (SIVVA), from the São Paulo City Council Health Department.
The Report/Case Description is written in Comic Sans MS, a friendlier font than Calibri, Arial or Verdana, with the intention of supporting the victim in her decision to go to the nearest DDM and allaying her sense of fear of returning to this institution. In contrast, the other forms are survey-like questionnaires designed by the data-receiving institutions, without implying any direct positive effect for the victim. The state and municipal notification forms use the World Health Organization's epidemiological categories for violence, distinguishing among assaults, accidents, injury inflicted on oneself, and injury by external causes (Schraiber, D'Oliveira, and Couto 2006).
The multiplication of bureaucratic forms at the face-to-face level of assistance can be interpreted as a strategy for "minor bureaucrats' survival" (Herzfeld 1992, 121), or as the effect of a gradual enlargement of the institution's connectivity, "wanting to be part of the game" (Lorenzo 2010;Moore 2000). Ironically, the larger the number of institutions that become part of the assistance network, the more reference to external institutions becomes an answer in itself and the more bureaucratic forms there are to fill out. Popular legal advocates do not follow up on the cases afterward. As one of them revealed, "We want to know about them, but we need to have a structure for that…. Sometimes we call the women's center…. We cannot call them at home, they don't want to speak in front of their relatives. It is very dangerous."

Final Remarks
We have attempted to develop an ethnographic comparison of apparently disparate institutional responses to victims of domestic violence. We chose to approach the responses of professionals from a Geertzian perspective that contrasts two ways of knowing, one consisting of the skeletonization of facts to law, the other of the schematization of social action. Both cognitive methods orient practitioners' legal sensibilities toward two different ways of understanding and interpreting the events narrated by the victims, and, therefore, of adjudicating truth and realness to the empirical chaos of first testimonies.
However, we conclude that while Geertz's interpretive relativism sheds light on the understanding of truth adjudication in particular situations, the two ways of knowing contrasted leave little room for understanding the actual replies that deviate from what would be expected from each ideal situation: skeletonization, as typical of legal/police officials, and schematization, as typical of human rights practitioners.
Indeed, ethnographic evidence revealed an important irony-the victims also skeletonize the facts that happened to them with the goal of making them fit their purposes, a phenomenon registered by countless studies on DDMs. Likewise, police officers also schematize social action, by associating certain perpetrator profiles and victims' social conduct stereotypes to specific types of crime. As pointed out in prior research on special civil courts' lawsuits, family stereotypes created or endorsed by judges influenced how homicide cases were evaluated more than the criminal facts being judged (Ardaillon and Debert 1987;Corrêa 1983).
We found that, more often than not, verbal offense and life threat are filed by police clerks. Particularly, life threats stand out as some of the most serious forms of violence suffered by victims attended at DDMs. As has been frequently registered, many women go to a DDM for the first time after having been threatened by partners who keep a gun at home. However, they reconstruct the story to save their male partners from hard penalization, such as imprisonment, as the case analyzed above showed. This phenomenon has not been sufficiently researched.
Criticisms apart, by asking in what ways legal factualism, reductive information-giving models, or the nonaccomplishment of empathetic expectations can be more harmful than helpful to victims, we should be able to push our analysis beyond dichotomous thinking. In observing each institutional setting more closely, it becomes evident that at the hospital, listening without time limits turns out to be an important challenge in a rushed and conflictive environment like the emergency ward. At DDMs, on the other hand, the rapid and fragmented way the information is conveyed to the victims makes it difficult for them to understand the law and to get an idea of what resources are actually available to them. A pre-attendance service is necessary. Women spend many hours in waiting rooms, sometimes an entire day. An informative video could be shown or someone could talk to them and answer questions, among other helpful activities.
Despite the institutional constraints observed in each case, both professionals and victims enjoy a certain margin of choice, though limited, in interpreting incidents and categorizing them. In the end, as a vulnerable group, female victims find themselves in a differentiated position, though one that is highly restricted by family, legal, and institutional constraints.