This document condenses the exchanges and dialogues we had during the process of translating Data Feminism and reflects the theoretical and semantic decisions that were made. This document is divided into two parts. The first section describes general issues about style, the challenges of translating from English into Spanish, and a series of conceptual recommendations on issues that, from Latin America and in Spanish, we interpret and experience differently, and therefore name differently. In the second section, specific matters of translation and interpretation are discussed.
This new version of Data Feminism is a collective and activist translation from a decolonial feminist perspective situated in Latin America. It seeks to reflect the spirit of the authors and to bring it into dialogue with the discussions taking place in our region.
The work was led by DataGénero, Observatorio de Datos con Perspectiva de Género. Special thanks to Catherine D'Ignazio and Lauren Klein for their confidence in this project.
Coordination and editing: Mailén García
Translators: Gina Ballaben, Sofía García and Ivana Feldfeber
Contributors: Helena Suárez Val, Diana Duarte Salinas, Lucía Peyrano, Florencia Aguilar, Giselle Arena, Ana Amelia Letelier and Patricia Michelle Garcia Iruegas.
Buenos Aires, March 3, 2023.
General aspects
"Words do not belong to writers, words belong to those who desire them, and in the history of the West's linguistic plundering, we have been taught to give up that desire, to give up our undisciplined, rustic, maroon words.”1
On the non-sexist use of language
Language influences representations and imaginaries about social life, the construction of prejudices, stereotypes, and the unequal distribution of privileges and vulnerabilities. Due to its binary grammatical gender, feminine and masculine, the Spanish language is prone to sexist usage. The generic employment of the masculine gender has invisivilized both women and the diversity of identities that make up our societies.
It is commonly thought that power relations are rooted in natural phenomena, however they are crystallized through a sexist use of language. It is therefore necessary and possible to incorporate a perspective towards equality that concludes the unequal connotations that have been given to one grammatical gender over another. For instance, the expression hombre público (public man) refers to a male who is active in politics, as opposed to the expression mujer pública (public woman) , which has been associated with a woman in prostitution (as it appears in the Diccionario de la Real Academia Española).
As Teresa Meana, a Spanish philologist who has promoted the non-sexist use of the Spanish language since the 1990s as a tool for feminist transformation, argues:
Language is, at the very least, a reflection of reality, of the society that uses it. As well as society, language is also racist, classist, heterosexist. And, of course, it reflects the inequalities derived from the situation of discrimination against women and reflects all the existing sexism and androcentrism. And since both androcentrism and sexism are manifested in the different uses of language (n/d:11).2
The present translation has favored the use, whenever possible, of gender-neutral expressions in Spanish. In the cases where it was not possible, a non-binary term was chosen, using the feminine ("a"), the masculine ("o") and the neuter ("e") as agreed between the authors of the book and the team of translators involved in this project.
This translation follows the criterion taken from a non-sexist use of language by utilizing the word "males" instead of "men". Historically, the generic masculine was the word used to refer to all of humanity in general. It is this androcentric use of language, where men seem to be the maximum reference and authority, that we want to avoid in this translation.
The sexist use of language makes linguistic inequality evident, which the sociologist Pierre Bourdieu designates as symbolic violence. For the author, symbolic violence is not exercised directly, nor is it physical, but it is evident in the asymmetrical schemes of power: in this particular case language. "What is not named does not exist," Steiner would say, and a position in line with our own.
To reflect the activism existent in the language, the decision to include the "e" as a non-binary expression arose from a consensus reached with the authors. As the translation process progressed, it became evident that many political slogans were weakened by the search for neutral expressions. It is not the same to say, "people working in science" as "scientists [feminine], scientists [masculine], and scientists [nonbinary]"3. Consequently, we decided to also include the gender-neutral as an option.
On translations and artificial intelligence
Artificial intelligence, so closely associated with working with and on data, appears to be a mystical option to solve problems on a large scale, faster than people, and "error-free". In line with the proposal of Data Feminism, these translator notes echoes the warnings about relying on automated translators online as a way of carrying out such work. While we recognize the potential of these tools, especially in terms of facilitating access to texts that would otherwise not be available in Spanish, we warn of the gender biases that they reproduce, given that they are still far from favoring a non-sexist use of language. This is why expert translators with an interest in this matter are essential and irreplaceable by digital tools.
During the work, and to verify that what the book narrates continues to happen and impacts our lives, especially those of women, LGBTI+ people, and racialized identities, we conducted an experiment. We selected random parts of the book and copied them into two automated translators to see the differences and similarities in the translation. Ironically, we found that "the researcher Timnit Gebru" was translated as "el investigador Timnit Gebru,"4 rendering invisible the gender of the Ethiopian-American researcher who works for diversity in technology and is co-founder of Black in AI and the DAIR institute dedicated to interdisciplinary research for decentralized Artificial Intelligence.5 We also noticed that "man" was systematically translated as "hombre" and that the pronoun used today in English for the gender-neutral indicator "they" (which also refers to the neutral third person plural, since in English this pronoun does not distinguish between male and female) is translated as "ellos", the pronoun for the masculine third person plural in Spanish. It is noticeable, that automated translators do not give the possibility of a neutral expression and force the masculine gender into a gender identity that does not correspond to it.
The translation process is neither neutral nor literal, each word has a meaning and, as Meana points out, language "reflects the inequalities derived from the situation of discrimination against women and reflects all the existing sexism and androcentrism" (n/d:11).6
For that reason, if you combine translations based on AI developments that were designed by elite, heterosexual, white, able-bodied, cisgender men from the Global North who don't even notice the "privilege hazard" with a language such as Spanish, which reproduces sexism, the result is "el investigador Timnit Gebru." The example is extremely illustrative. The AI translation makes invisible what Dr. Gebru has been working on for a long time. It is precisely these kinds of outcomes that Data Feminism seeks to point out and transform. The book deals with these issues. This is why it should be made clear that the translation was not automated but rather crafted collectively.
Translating with a decolonial feminist perspective
The next question we want to consider are the positionalities of those of us who work in the translation process. As Latin American women, feminist activists, and data specialists, we translate from an intersectional feminist approach and with a decolonial perspective. What does this mean? It means that we translated from this epistemic position, seeking to align our translation work with our values.
This stance and approach can be seen, for example, in the way the concept of "people of color" was translated as "racialized identities."7 In Latin America, use of the latter term is slowly spreading among grassroots organizations seeking to highlight the impact of colonial processes in the construction of the concept of race and related oppressions. The term “racialized” seeks to emphasize that race is a social construct imposed by dominant groups on oppressed groups. It also allows for an emphasis on race as an action that is exercised on a group of people, an act on one or more racial groups. It is important to highlight that each country or region has different racial regimes; therefore, racialized people in one part of the world are not necessarily the same as in another part of the world, since the people who hold hegemonic power vary.
The concept of "racialized bodies" has been developed from the fields of anthropology and postcolonial studies. Restrepo argues that "racialization can be considered as a particular constitutive marking of bodies" (2010:18).8 In "Descifrando nuestros cuerpos racializados", Karina Ochoa Muñoz describes the process of racialization, its implications, and consequences, as well as the differentiated sexuation among these racialized bodies. She argues that racialization involves the construction of historically minoritized bodies that arise from the configuration of "mechanisms and devices of social marking distinctive of the modern ethos" (2014:3).9 Therefore, it is a matter of thinking of the body as a territory and from there understanding the mechanisms of domination and oppression that are exercised over it and the subjectivities of all people.
Translating "people of color" as “racialized identities” is a way of showing that there are oppressions based on the concept of race that extend beyond skin color. At the same time, it is a category specific to the Global South, which is already racialized by the process of colonization it has suffered. It is also a term frequently used by postcolonial and decolonial feminists such as Karina Bidaseca and Karina Ochoa Muñoz.
DataGénero is an organization that aims to build a sustainable and inclusive data future from and for Latin America. In that process, we are building on multiple borders and boundaries.10 For instance, at the intersection between data and gender theories, between activism and academia, between civil society and the State, between reading the North and writing from the South, civil society and the State, in that of reading the North and writing from the South. This work is a result of this.
Yuderkis Espiñosa Muiñoso in "De por qué es necesario un feminismo descolonial: diferenciación, dominación co-constitutiva de la modernidad occidental y el fin de la política de identidad "11 argues that the fourth pillar of decolonial feminism is:
the contributions of some of the western feminist theories that, in their commitment to the critical revision of feminist theory, contribute methodologies and categories of analysis to which we decolonial feminists have recourse, generally after submitting them to revision. As problematic as these appropriations tend to be, they become useful in the critical revision of feminism and, when mixed or operating within an analytical framework such as that of decolonial feminism interested in the critique of Eurocentrism and epistemic racism, they can come to function in a different way, assembling a kind of intermediate stage in the task of decolonizing feminism (2012:152).12
This translation is a bid to start down a similar path with regard to data feminism. It is about having, in our language, a book from the Global North, written from the United States – to review it, appropriate it and discuss it in our terms. That is what we wanted to reflect on in this document. We have not yet managed to publish similar books in our languages.13 We hope that these chapters of Data Feminism will contribute to foster debates from activism and academia that will result in books from and for Latin America on data feminism.
The inequalities that structure the world are also reflected in both the data agendas and the ways in which they are thematized in the North and the South. Thus, while in the Global North there are more and more books warning about the problems caused by racial and gender bias, published by prestigious publishers, and freely available on the Internet, in the Global South there are blog posts and informal documents circulated by activists to warn about the issue.
Above all, this is an activist translation. It has been a year since DataGénero started a collective project of translation. We perceived the need to make these books available in Spanish. We want to "use" them to enhance our discussions and debates, especially in academic circles that are so reluctant to activist forms of knowledge construction. The first step we took was to write Catherine D'Ignazio a Twitter message in Spanglish to propose a translation of the book Data Feminism. In the summer of 2022, we translated the first few chapters. Then, Catherine and Lauren got the editorial permission to make the full translation that you can find available today. This is an activist collaboration between the authors and us, a feminist way of circulating books, ideas, and promoting struggles that are global, thoughthey take on different dimensions depending on their place and their context.
Specific issues
We have also reached other more technical agreements related to translation specificities, which are detailed below:
The titles of articles, books, and other references have been translated into Spanish. This will help those people who read the chapters to better locate the information, regardless of whether translations into Spanish exist. Consequently, we prioritize the fact that we want all Spanish-speaking people to know the titles of these references.
Black women were translated as "Mujeres negras". On this occasion, it was decided to maintain the capital letter to vindicate the collective, a frequent practice in English and not so much in Spanish.
Blackness: black community (comunidad negra).
People of color was translated as racialized identities (identidades racializadas).
EE.UU. is the acronym used to refer to the United States of America.
The names of USA Today, Glamour, New York Times, CNN, The Economist, ProPublica, Bloomberg and BusinessWeek were kept in English because that is how they are also known in Spanish. The same criterion was used for Human Rights Watch. It was also followed for the projects known as AI Now and IT for Change.
“Diversity in Faces” (DiF) was translated as "Diversidad en Rostros" and the Algorithmic Justice League as “Liga de Justicia Algorítmica”.
“Privilege hazard” was translated as "privilege risk" (riesgo de privilegio).
The concept of “who questions” was translated as "preguntas por el quien."
“Maternal harm” was translated as "obstetric violence" (violencia obstétrica) since there is no more inclusive term in Spanish to describe that phenomenon, which is not limited exclusively to the exercise of violence by health personnel.
“Native American” and “Indigenous” were translated as "original peoples" (pueblos originarios).
The term “genderqueer” was translated as "genderfluid person" (persona de género fluido).
The term "queer" was kept in English to refer to queer academics because in Spanish they are referred to in the same way. It is important to note that if people from the Global South are mentioned, they can be referred to as "cuir" (with a c)14.
Instead, when it referred to "queer people," it was translated as the LGBTIQ+ (lesbian, gay, bisexual, transgender,, intersex and queer) communities.
The term "nerd" was kept in English because it is frequently used in Spanish, there is even the term "nerdo/a" in the dictionary of the Real Academia Española.
“Ability” was translated as “capacitism” (capacitismo). Capacitism is a form of discrimination or social prejudice against people with disabilities.
“Minoritized groups” was translated as "minoritized groups" (grupos minorizados). This is a concept used in the book to describe groups of people who position themselves in opposition to a more powerful social group. While the term “minority” describes a social group composed of fewer people, minoritized indicates that a social group is actively devalued and oppressed by a dominant group, which has more economic, social, and political power. For instance, regarding gender, men constitute the dominant group, while the other genders constitute minority groups.
With respect to the term “big data” a double standard was used: when referring to technical process it was translated as “big data”. For example: "questions about big data versus micro data or quantitative versus qualitative data are too often framed as false binaries". When the concept referred to the theoretical construction that exists about them, it was left in English, understanding that it is an anglicism that is usually not translated.
Creative Commons Attribution 4.0 International License (CC-BY 4.0)
