Is it the same if we say that Mary was raped, or that James raped her? No. The words we choose to tell a violence are not neutral as they contribute to shape the perception of readers or listeners about the event, its context, the victim and the aggressor.
A lot has been written about gender-based violence and its representation. In Italy, one of the most interesting recent works is the STEP project by Università degli Studi della Tuscia and professor Flaminia Saccà, in partnership with Associazione Differenza Donna. STEP researched bias and stereotypes about violence against women through an accurate sociolinguistic analysis on more than 16,700 articles published between 2017 and 2019 by 15 national and local media, and about 280 court verdicts.
The research showed a distance between the real number of gender crimes and their media coverage. According to the Ministry of the Interior, in 2019 the most reported crimes dealt with home violence (51% of total occurrences), stalking (31%), sexual assaults (17%), feminicides (less than 1%), and slave trading (less than 1%). But media focused on stalking and feminicides, generating 53% and 44% of the total articles respectively, followed by home violence (14%) and sexual assaults (10%).
“Media select the events that become news (…) Articles should not be taken as the exact count of reality, but this social representation is indeed insightful,” wrote professor Saccà. “Home violence is not a story, seems like it is almost normal. But normalizing this kind of violence leaves women alone and helpless.”
STEP also highlighted a misrepresentation of gender violence, with lots of bias and stereotypes that blame women and tend to excuse aggressors. Both in media and court outlets, STEP found several occurrences where the female victim is described as responsible for the violence she suffered (so called ‘victim blaming’, for instance underlining that she was dressed inappropriately, was drunk, or didn’t oppose with enough determination), or the aggressor’s viewpoint is favored (so called ‘himpathy’), presenting the violent act as an excessive but understandable reaction to a certain behavior of the victim.
Gender violence is often neutralized using two apparently different techniques: either it is normalized, presented as the result of family conflicts or failed relationships, or it is reported as something exceptional, the consequence of an uncontrollable burst, or the act of a deviant subject. The outcome is the same, that is making the man and his accountability vanish.
“The still dominating patriarchal culture continues to make female accusation a taboo, and picture male violence as something justifiable. This normalization process ends in legitimizing gender violence, and reiterating it,” continued Saccà.
There are specific rules for the correct representation of gender violence in the Istanbul Convention (2011) and the Venice Manifesto (2017). In 2021, the Act of journalists’ obligations (‘Testo unico dei doveri del giornalista’) reaffirmed that, when reporting gender violence, media should always avoid stereotypes, select respectful words and images, be factual, and consider the impact on families and relatives.
Although media attention is growing, these principles are not always applied. So, how can we fight stereotypes and bias that, consciously or not, compromise the representation of gender violence?
We need a deep change in mainstream culture. We need powerful learning programs in schools. We need trainings: the above-mentioned STEP project engaged about 2,000 professionals including judges, lawyers, police officers, and journalists with workshops and lessons.
Technology can also help. A team of researchers from Pavia and Groningen universities developed an Artificial Intelligence algorithm to scrutinize texts about feminicides and gender violence and predict possible reactions by receivers. For instance, the system can tell if the aggressor’s responsibility is clearly stated, if there is any ambiguity, if the description of the victim is biased. The goal is to launch an upgraded version that, starting from a sentence and its predicted interpretation, can suggest correct and respectful alternatives.
