Making femicide visible

Inaudibility and invisibility

After the women’s revolution of the 1960s, sociologists began studying gender and realized that women were not only under-studied or not counted, but they were ‘inaudible’. Edwin Ardener (1975) suggested that women, as a group, were muted. Their voices were either unheard, ignored or silenced. The concept of inaudibility developed into muted group theory, a critical theory explaining that whole groups of marginalized people were muted (Ardener and Ardener, 2005: 50–54). Women were targeted, despite the fact that they represented half the human population. It had already been substantiated that women and men developed different linguistic perceptual grids, with differing perspectives and interests based on gender (Kramarae, 1981; Weil, 1983). Muted group theory took that idea one step further, elaborating how asymmetry in hierarchical relations affects communication between dominant holders of power and subordinated, silenced groups. Kramarae proposed that linguistic communication was initiated and controlled by men, who retained domination over powerless and inaudible women, and that the theory is as pertinent today as when it was conceived (Kramarae, 2005). Brescoll (2011) confirmed that women’s volubility is restricted in the organizational workplace.

Although the inarticulateness of women continues, both in the field and in scholarly works, which often still portray entire societies from an androcentric point of view only, much has been done to remedy the situation. Women’s histories and attitudes are now emerging in sociological literature. However, while women’s inaudibility has been documented, some gender-related phenomena remain inexcusably invisible. The purpose of this article is to bring femicide to light.

Ralph Ellison’s Invisible Man (1952) was a milestone, both on the literary scene and for sociology. Laypersons and scholars alike, who delved into the question of unperceivable and unobservable black Americans in the dominant white ‘host’ society, took up the theme of invisibility. Not only were these citizens covert, they were imperceptible. In 2013, Wingfield returned to the concept of invisibility, a theme on which she had published in the past (McDonald and Wingfield, 2008). In her study, she interviewed 42 upper middle-class black professional men: physicians, lawyers, engineers and bankers. She found that the concept of visibility was applicable in that professional men continued to be ‘invisible’, while the tendency in society and among academics is to emphasize either black men who dropped out, or exceptional, elite black men, such as Barack Obama. Hand in hand, Wingfield found that the same professional men were gradually becoming more visible through workplace tokenization, but this highlighted their difference from a normative white, male worker. Today, some would interpret tokenization as affirmative action and point to the similarities and differences. While tokenization subjected the black professionals to extra scrutiny and pointed out that they were exceptions, the outcome was far more complex. In effect, the men were ‘partially tokenized’ (Wingfield, 2013), being marginal and disadvantaged, yet privileged at the same time. In brief, Wingfield used the concept of invisibility, like inaudibility, to document a group of people.

While volubility can only be used heuristically in relation to groups of people, I would argue that the metaphor of visibility can also be applied to social facts that sociologists do or do not see. Social facts are the effect or creation of human activities, actions or agency. They are not the product of conscious intentions, but the unanticipated consequences of human behaviour or agency. According to Durkheim (1982 [1895]), a social fact is any way of acting, whether fixed or not, which is capable of exerting over the individual an external constraint. A social fact is perceptible in the whole society, while having an existence of its own, independent of its individual manifestations. Examples of Durkheimian social facts are social institutions, such as kinship and marriage, political and economic organization, and religion. In his analysis of suicide, Durkheim (1951 [1897]) showed that an individual act of suicide is part of a suicide rate in an individual’s social group, which is a social fact and not an individual disposition. By the same token, another social fact is homicide, comparable indeed to suicide. One type of homicide, femicide, or the murder of women because of their gender, is a social fact relegated to invisibility for all these years, which is only now emerging as visible as a new social consciousness about women’s dignity and equality develops. In the case of suicide, social currents are expressed as suicide rates, which differ among different groups in society. These rates show regularities over time, with changes in the rates often occurring at similar times in different societies. The rates can be said to be social facts in the sense that they are not individual, but are societal characteristics (Durkheim, 1951 [1897]: 48, 51).

Durkheim’s first task in tackling suicide as a social fact was to define it. He selected the following definition: ‘Suicide is applied to all cases of death resulting directly or indirectly from a positive or negative act of the victim himself, which he knows will produce this result’ (Durkheim, 1951 [1897]: 44). This included death as the result of a self-inflicted act, like cutting oneself or shooting oneself, or a negative act, like fasting till death with intention to die. He then went on to identify different types of suicide. The intention in this article is neither to dwell upon Durkheim’s analysis of suicide, which has been the object of debate for well over a century (Pickering and Walford, 2000), nor to discuss the connection between femicide and suicide, which has been studied by different researchers, according to various theories. Often, the will to separate in intimate relationships is the catalyst for femicide and the subsequent suicide of the killer (Ellis et al., 2015: 99–125). Rather, the point of this article is to show that before one can discuss femicide as a social fact, one has to define it, bringing into question the varying types, and then understand it.

Definitions of femicide

Originally, Russell, who coined the term ‘femicide’ at the International Tribunal on Crimes against Women in 1976, claimed that it was an act motivated by a patriarchal and misogynist culture (Radford and Russell, 1992: 3). In 2001, she redefined the term to refer to the killing of females by males because they are female (Russell and Harmes, 2001).

In recent years, utilization of the designation ‘femicide’ has become the consensus, although ‘feminicide’ is certainly more acceptable in the Spanish-speaking world. Lagarde, in the ‘Introduction’ to the Spanish version of Radford and Russell (Lagarde, 2006), refers to ‘feminicidio’ in a global perspective, although it is interesting that even she has championed the use of the word ‘femicide’ in Central America (Shulman, 2010). ‘Uxoricide’ and other terms appear to be dying out, although they still exist in the scientific literature as late as 2011 (see Mize et al., 2011). While the term ‘femicide’ was already current in North America (cf. Campbell and Runyan, 1998), the first reference to femicide in the European Parliament was as late as 2006, during the course of a hearing on ‘feminicide’ in Mexico and Guatemala. In 2007 the European Parliament adopted a resolution on Mexico and Central America, committing themselves to combat the murder of women, still designated as ‘feminicide’ (Dimitrijevic et al., 2015: 56). In November 2012, the Vienna Declaration on Femicide in the United Nations in Vienna, in conjunction with many organizations including UNODC and the UN Commission on Crime Prevention and Criminal Justice, proposed a broad definition:

… that femicide is the killing of women and girls because of their gender, which can take the form of, inter alia: (1) the murder of women as a result of intimate partner violence; (2) the torture and misogynist slaying of women; (3) killing of women and girls in the name of ‘honor’; (4) targeted killing of women and girls in the context of armed conflict; (5) dowry-related killings of women; (6) killing of women and girls because of their sexual orientation and gender identity; (7) the killing of aboriginal and indigenous women and girls because of their gender; (8) female infanticide and gender-based sex selection foeticide; (9) genital mutilation related deaths; (10) accusations of witchcraft; and (11) other femicides connected with gangs, organized crime, drug dealers, human trafficking and the proliferation of small arms. (Laurent et al., 2013: 4)

The General Assembly of the UNODC, in its Agreed Conclusions adopted in the 57th session in 2014, refers to femicide as ‘gender-related killing of women and girls’.

As with suicide, femicide can be distinguished according to type. It includes so-called ‘honour’ femicides, sex selection before birth, dowry marriage femicides and a host of other manifestations of extreme violence culminating in the death of a woman. Intimate femicide is just one form of femicide perpetrated by a familiar person, usually a family member. It includes murder by intimate partners and killings which occur when a woman is killed by a male family member for dishonouring the family status (Gill et al., 2014; Sev’er, 2013). The term ‘honour killings’ has been criticized by some scholars (e.g. Shalhoub-Kervorkian and Daher-Nashif, 2013), who prefer to regard these kind of murders simply as ‘femicides’, which should be examined in the wider context of colonization. Intimate partner femicide is the final act of domestic violence or intimate partner violence, and is often the ultimate result of years of suffered violence. A recent study affirms that 39% of all femicides (and 6% of all homicides) are intimate partner murders; in high-income countries, the percentage rises to 41% of all femicides (Stöckl et al., 2013).

As with suicide, rates of femicide vary from year to year and from country to country. Differing rates across regions and cross-national variations have been reported widely (Corradi and Stöckl, 2014). During the period 1985–2010, female homicide victimization (a designation Stamatel chooses to use instead of the term ‘femicide’) increased in some countries in Europe (e.g. Switzerland, Slovenia and Portugal), remained relatively stable in others (e.g. France and Italy), while countries such as Norway had extremely low rates of femicide. Accounting for macro-level variations in female homicide victimization requires knowledge of socio-political trends, such as post-communism, as well as an understanding of different criminological theories (Stamatel, 2014).

Reasons for invisibility

Compared to domestic violence or homicide, femicide per se has been studied sociologically in limited contexts only. This section postulates reasons for the invisibility or submersion of the phenomenon.

However, as sociologists, we often study insalubrious subjects, such as slavery, war, violence, oppression and abject social differentiation. Violence is at the centre of sociological enquiry (see Collins, 2007). Stallings (2002) developed methodological tools for the study of natural disasters. Skjelsbæk (2007) wrote about the aftermath of the Bosnia-Herzegovinian war rapes. It is not for us to decide whether femicide is worse or better than these other atrocities; they are all social facts worthy of sociological study.

Femicide has indeed been given limited attention in other disciplines, but even less in sociology. As mentioned above, it is often called ‘female homicide victimization’ or the like, thereby evading the political agenda of the designation.

I would argue that femicide is the very stuff of sociology. It is akin to suicide and can be analysed as a social fact utilizing Durkheimian and post-Durkheimian conceptual tools. We should take into consideration the goals of sociology as laid out in classic texts, such as that authored by Ginsberg (1927). In ‘The Scope of Sociology’, he defined sociology as ‘the study of human interactions, their conditions and consequences’. He continued that sociology ‘should deal with the whole tissue or web of social relationships’ (Ginsberg, 1927: 135). The study of femicide, whether perpetrated consciously as an act of will or unconsciously or irrationally, falls squarely within the realm of sociology.

Traditionally, males have dominated sociology. The great giants of the last 100 years from Max Weber to Bruno Latour all devalued women’s contributions. Eisenstadt in the comparative method, Bauman in post-modernist theory (Weil, 2011) and Duncan in social measurement all ignored gender as a major thrust of sociology: femicide was certainly not considered a major sociological issue. Men continue to dictate sociological thinking at the upper echelons of the discipline. For men, it is possible that femicide is conceived as a subject to be studied by women only, and therefore unworthy of study.

The feminist theoretical standpoint has emerged as a major thrust in sociology that has significantly reshaped the discipline and radicalized sociology (Harding, 2004). By calling attention at the micro-level to the powerful impact of gender in the social ordering of relationships, as well as pursuing macro-level analysis of gender in institutions, feminist sociology set a new agenda. Yet feminist sociologists, who were quick to take up most issues pertaining to gender, still ignored femicide with only a few, lone exceptions.

By contrast, I would maintain that femicide is a legitimate subject to study, both for self-defined feminists and for sociologists in general. Femicide need not be relegated to the confines of radical feminist sociology alone.

The proponents of these ideas insist that femicide cannot be studied per se, but must be linked to other social phenomena. I will dwell upon one of these ideas, namely, that femicide is genocide. Banerji (2009) has proposed that ‘female genocide’, as she calls it, is one of the three most pressing issues in India and the manifestation of sexual malfunctioning of Indian society. She also launched the ‘Female Genocide in India and the 50 Million Missing Campaign’, in which she promoted the idea that India represents countries where there are millions of ‘missing’ women and girls, some killed off before they are born.

At the COST annual conference in Zaragoza, which took place in March 2015, Russell, who had originally proposed a misogynist cause for the murder of women, now suggested that femicide is genocide. In a Skype lecture entitled ‘International Mass Femicide: The Most Extreme Form of Genocide’, she said: ‘I agree with Banerji that these gigantic numbers of femicides must be recognized as cases of female genocide. Six million Jews were exterminated during the Nazi Holocaust, compared to 50 million femicides in just one country today.’ Russell claimed that femicide is the most prevalent form of genocide, and included different manifestations, including witchcraft, denial of reproductive rights for females, AIDS as mass femicide and rape as mass murder.

The idea that femicide is genocide is not new. However, I would maintain that femicide cannot be defined as genocide or equated with it. Genocide is the mass murder of a people, like the murder of over 1 million Armenians during and after the First World War, or the murder of 6 million Jews in the Holocaust during the Second World War. Females do not constitute a genus, and a genus cannot continue without females.

Qualitative studies of femicide are scarce because the victim, who would be the object of study, has been eliminated.

However, methodological solutions do exist. Researchers can focus upon women who were nearly killed, that is, women who experienced extreme acts of violence and/or survivors of what I term ‘failed femicide’ attempts. They can study the victims’ kin, friends or neighbours. Alternatively, they can study perpetrators.

The number of women who survive femicide attacks is small. Moreover, it is not easy to interview the survivors. In some cases, the victim is physically in such bad shape that she is unable to talk. In most cases, the judiciary and the police ‘guard’ the survivor, isolating her from her own environment, and certainly from researchers. The inaccessibility of ‘failed femicide’ survivors is even more acute among migrants with whom western researchers (of whatever origin) under normal circumstances have little interaction.

One exceptional study interviewed 30 women aged 17–54 who had survived an attempted homicide by an intimate partner, with in-depth interviews in six cities as part of an 11-city case-control study to determine the risk factors of actual and attempted intimate partner femicide (Nicolaidis et al., 2003). Victims participated in an audiotaped, semi-structured, in-depth interview of 30–90 minutes’ duration. ‘The purpose of the interview was to allow women to describe, in their own words, their relationship with the partner who had attempted to kill them, and their perceptions of the activities and events that preceded the attempt’ (Nicolaidis et al., 2003: 780).

An alternative is to interview family members, friends, neighbours or social workers and health care personnel who were in close touch with the victim prior to the event. In many cases, family members, frequently children or mothers-in-law, are witnesses to the act of femicide (Dobash and Dobash, 2012).

Most of the qualitative literature on intimate partner homicide bases itself on interviews with perpetrators (cf. Shanaaz et al., 2011). Polk (1994) published a rich criminological study of homicide by men of men and by men of women, with a particularly large number of case studies (100).

Data on femicides are provided by multiple sources, most frequently from official sources like the police, but also from internet sites of different (usually feminist) organizations, court files and ministerial offices. In some countries, these statistics are not made available to the public. In order to access them, one needs a lot of patience and pull: each data source has its own agenda.

Across the board, missing femicide data is a rampant problem, either because the state did not want to collect data on the phenomenon, or the police closed the file due to lack of evidence, or due to unreported cases. In official statistics, femicide is usually subsumed under other types of murder or homicide and therefore difficult to isolate.

Official statistics which governments finance, produce and turn into the basis of decision-making are the object of much sociological debate. Some categories are included, others are excluded, often linked to political decisions and climates. Official statistics are, after all, human constructions; they show impartiality and rigour but are historically mobilized by governments for their political uses (Desrosieres, 2002). They illustrate what is important and what is not for a particular society or era in history. Following Duncan (1984), it can be shown that social and economic statistics share common features with other forms of measurement in that they serve to control society (Starr, 1987). The fact that the vast majority of countries do not have official statistics on femicide, as opposed to homicide or other forms of murder, is testimony to its invisibility. Intimate partner femicide, in particular, is difficult to document. In some countries, there is no obligation to report the gender of the victim in national homicide data. In addition, it is sometimes difficult to determine that a murder was actually a femicide, when the relationship between the victim and the perpetrator may be unknown. Nor do the media necessarily report the murder with the name of the victim or perpetrator (such that one cannot know if the murdered person is female); in fact, they may not report it at all. In 2013, Stöckl et al. found that information was lacking in 21% of all cases of homicide on the relationship between the victim and the perpetrator.

A way round official statistics is to trace femicide cases through newspaper archives and media reports, but this in turn has its limitations and reflects the public and media’s view of the lethal killings of women.

Quantitative studies could also include surveys to determine the causative factors of femicide, without necessarily focusing on macro-statistics or mega-data. McFarlane and colleagues studied 141 femicide cases and 65 attempted femicide survivors, in order to examine the phenomenon of stalking prior to an attack (McFarlane et al., 1999). The data were gathered in 10 US cities during the period 1994–1998. The victims were identified from closed police records and contacted by mail. Once they consented to be interviewed, trained doctoral students ran a questionnaire, including an 18-item stalking survey; the interview took one hour. Campbell et al. (2003) carried out a large survey of proxies of 220 intimate partner femicide victims identified from police or medical examiner records, along with 343 abused control women. The researchers concluded that pre-incident risk factors included the perpetrator’s access to a gun or a previous threat with a weapon, the perpetrator’s stepchild residing in the home and estrangement, especially from a controlling partner.

A more salient answer concerning the measurement of social phenomena could be made if masculinity models in society were to be reduced. In patriarchal societies, interest in women and women’s plight is low and this is reflected in the categories selected when devising official statistics. Since information is absent on the rates and incidence of femicide in most countries, this is just another means of controlling women. In order to produce reliable statistics, the attitude towards women in general and femicide in particular must change.

In view of the large quantities of missing data and incomplete official statistics on femicide, it is clearly difficult to measure rates and then compare them. The comparative method is one of the backbones of sociology, championed by all its great masters, including Durkheim. Durkheim argued that all sociological research is essentially comparative since social phenomena are always held to be typical, representative or unique, and these of themselves imply some sort of comparison (Durkheim, 1951 [1897]). In today’s terms, sociology has to be cross-cultural and transnational.

Comparing femicide rates therefore remains the ultimate challenge for femicide researchers. Scoring methods and the production of official statistics cross-nationally will have to be coordinated. The comparison of femicide scores between different populations assumes that femicide has similar properties across samples, and the identical meaning is attributed to the phenomenon. Notwithstanding, in the past couple of years, scholars have begun to coordinate comparative cross-national surveys of femicide based on available data (e.g. Corradi and Stöckl, 2014; Stamatel, 2014; Stöckl et al., 2013), and other comparative studies are in the making.

Enhancing future visibility of femicide

None of the hypotheses could account for the dearth of scientific literature on the subject. Given the relative invisibility of femicide studies in sociological literature, until activists took up cudgels in the last two to three years, and given the concomitant difficulties in studying the phenomenon for different reasons by either quantitative or qualitative means, the question remains: How can femicide become more visible in the future?

Conclusions

In this article, I examined a hitherto invisible or quasi-invisible phenomenon, femicide. The article set out to illustrate that, while women have been shown to be muted and sociological theory has been devoted to the study of gender, femicide has not received similar attention. In 1982, Carol Gilligan published her then revolutionary book In a Different Voice. From that date on, thousands of articles, particularly by feminists, wrote of ‘giving women a voice’. With respect to femicide, the women killed cannot provide a voice. The situation is even graver in that the phenomenon cannot be vocalized by those involved, and is often invisible to others, including researchers.

Referring to femicide as a Durkheimian social fact like suicide, the article reviewed some definitions and types of the phenomenon. A series of hypotheses were then posited to suggest reasons why femicide has been invisible in sociology. None of the hypotheses could account for the dearth of scientific literature on the subject. The article suggested ways of enhancing future visibility of the phenomenon. Femicide rates are social facts, like suicide rates, and other social institutions and relations that humans create or enact.

As Ralph Ellison argued, invisibility is not the result of a particular condition, but the result of other people refusing to see him. As he wrote early on in the novel: ‘Well I was and yet I was invisible, that was the fundamental contradiction. I was and yet I was unseen.’ Upon receiving multiple awards, it was time for Ellison, the person, to emerge from the underground.

It is now time for femicide, a rampant phenomenon affecting thousands of women each year, which other people refrain from seeing, to come out of the sociological wraps and become a visible social fact.