repressed informatics

Compiled from: An article in Logic Magazine " Informatics of the Oppressed " (August 31, 2020) by Rodrigo Ochigame

Subtitle: Oppressive algorithms have been around for a long time. So do radical projects to dismantle these algorithms and build liberating alternatives.

How did you read this article? If you read online, you probably clicked on a link in an algorithmically generated recommendation or a list of search results. It could also be sent to you by a friend after finding a link in an algorithmically generated list. Whatever chain of operations brought you here, it likely involves an information retrieval system.

Such a system selects a few from billions of possibilities. Their existence is inevitable: at any one time, we can only understand a small part of a vast world. The problem is, the systems that filter the world are not designed for your benefit, but for the benefit of business. No word better describes the dominant form of information consumption on the Internet than "feed." Just like in animal husbandry, your information diet is designed to maximize the output of your business operations.

If Silicon Valley's PR department declares that their product is simply "finding the most relevant, useful results in an extremely short amount of time, presented in a way that helps you find what you're looking for," which is what Google says on its website A description of its search algorithm will be clearer to executives and shareholders. They know that the whole point of this industry is to make money for showing you what you don't need: advertising. They know that the whole point of this industry is to make money for showing you what you don't need: advertising.

For designers of commercial search engines, the conflict of interest between advertisers and users has always been obvious. In 1998, a few months before Google was founded, graduate students Sergey Brin and Lawrence Page presented their prototype web search engine at an academic conference. In an appendix to their paper, they comment, "We expect that ad-funded search engines will be inherently biased towards advertisers and away from consumer demand." Indeed. More than 20 years after this prophecy, all major search engines, first and foremost Google, now operate precisely on the business model of surveillance-driven targeted advertising.

The algorithms of these search engines are optimized for profit. The advertising industry controls most of the research and development in the field of information retrieval. Computer scientists and engineers often measure the "relevance" of potential results and test the "performance" of candidate algorithms against evaluation benchmarks and validation datasets dictated by industry priorities. The dominant system is designed to maximize ad revenue and "engagement" metrics like "click-through rate". Therefore, these systems tend to promote content that is already popular or similar to what the user has seen or liked before. Whether based on simple correlation and regression analysis or complex machine learning models, predictions of popularity and similarity tend to be predictable and similar.

No wonder the public sphere appears so impoverished in the digital age. The systems that govern the flow of political speech are often originally designed to sell consumer goods. This fact has major implications. Recent research has documented the disastrous effects of “surveillance capitalism,” particularly how commercial search engines deploy “algorithms of oppression.” These patterns of oppression of the oppressed are so common in the world that it seems impossible to design a system that does not repeat them.

But other options are also possible. In fact, starting in the 1960s informatics—information science—as an institutionalized field, anti-capitalists have tried to imagine less oppressive, and perhaps even liberating, ways of indexing and searching for information Way. In particular, two Latin American social movements—Cuban socialism and liberation theology—inspired experiments with different approaches to informatics from the 1960s to the 1980s. Taken together, these two historical moments can help us imagine new ways of organizing information, threatening the capitalist status quo and, most importantly, facilitating the widespread spread of ideas of the oppressed.

Struggle on the Library Front

What happens the day after the revolution? One of the answers is the reorganization of the library. In 1919, Lenin signed a resolution calling on the People's Commissariat of Enlightenment "to take immediate and most powerful measures, first to centralize library affairs in Russia, and secondly to introduce the Swiss-American system." Lenin was presumably referring to the The organization of European libraries he observed during his exile from Russia in the early 20th century. By imitating the "Swiss-American system", the Bolshevik leader wanted a single state system with centralized control over the distribution of books and the development of collections.

Forty years later, Cuban revolutionaries also recognized the importance of the struggles that Soviet leaders like Nadezhda Krupskaya once called "on the library front." sex. After the Cuban Revolution of 1959, Fidel Castro appointed librarian María Teresa Freyre de Andrade as Jose Martí of Havana The new director of the Jose Martí National Library. As a lesbian and longtime dissident exiled and imprisoned by the previous regime, she has long been concerned with library politics. In the 1940s, she put forward her vision for a popular library, a "popular library", different from a mere "public" library. A public library may be a "rather passive" library, where "the book sits quietly on the shelf, waiting for readers to find it", while a "popular library" is "very passive" "eminently active" because it "makes extensive use of propaganda and uses different procedures to mobilize the book and make it go." in search of the reader)

After the revolution, María Teresa Freyre de Andrade and her staff set out to implement this vision. They brought books to people by dispatching buses, bibliobúses, which acted as mobile libraries, to rural areas without libraries. They also began to develop new practices that revolutionized library work. Unlike Lenin, the goal was not to imitate the organization of European libraries. In a 1964 speech, Freire de Andrade argued that Cubans could not simply "copy what the British did in libraries". By doing this, "we will have a magnificent library, we will catalog it well, we will serve many people well, but we will not be actively involved in what is a revolution."

How can librarians actively participate in the revolution? One solution is to collect and index materials that were excluded from library collections or suppressed in the pre-revolutionary period, such as the publications of the secret revolutionary press of the 1950s. But librarians are also involved in a broader revolutionary project: Cuba's efforts to build its own computing industry and information infrastructure. This project culminated in a unique new field of information science, inheriting the revolutionary ideals of Cuban librarianship.

Redistribution of Information Wealth

Both the revolutionaries and their enemies recognized that information technology would be the strategic focus of the new Cuba. A former IBM executive recalled that “all foreign companies were nationalized except IBM Cuba,” because “the Castro government and most of the nationalized companies were users of IBM equipment and services. "But from 1961 to 1962, IBM closed its Cuban branch, and the U.S. government imposed a trade embargo that prevented Cuba from getting computer equipment. This meant that Cuba would be forced to develop its own computer industry with the help of other socialist countries in the Soviet-led Mutual Economic Assistance.

Between 1969 and 1970, a team from the University of Havana created a prototype of a digital computer, the CID-201, and an assembly language called LEAL, short for "Lenguaje Algorítmico" (language of algorithms), which is also " Loyalty means "loyal". The design of the CID-201 is based on the schematics in the manual for the PDP-1, a computer built by the Digital Equipment Corporation. Due to the trade embargo imposed by the United States, the team was unable to purchase the necessary electronic components in Europe, but eventually managed to ship the components from Japan into more than 10 official documents with the help of a Japanese-Cuban businessman in Tokyo. Bag.

Cuban mathematicians also programmed a chess-playing computer in the LEAL language; the computer even lost a game against Fidel Castro, recalls an engineer at CID-201 . Starting in the 1970s, Cuba built thousands of digital computers and even exported some computer parts to other CEA countries.

The rise of digital computing transformed library business in Cuba. Freire de Andrade welcomed the digital age, citing Marx and Engels, comparing computers to communism: "A ghost is wandering in the world of information, and this ghost is the computer; to our delight, This situation has started to change our [library] field, giving us a challenge to make [the field] more interesting than it is.” Cubans learned informatics techniques mainly through Soviet textbooks translated into Spanish. They combine the computational methods they will learn from these books with the revolutionary ideals of Cuban librarianship. This synthesis produced a theory and practice that was very different from Western and Soviet informatics.

Consider the concept of "information laws," a staple of informatics textbooks. A classic example is "Lotka's Law", proposed in 1926 by Alfred J. Lotka, a statistician from the Metropolitan Life Insurance Company of New York, who tried to The "frequency distribution of scientific productivity" is calculated by indexing the authors' publication counts in the abstract index of chemical publications. He claimed that this distribution obeyed the "inverse square law", which states that "the number of people who contribute twice is about a quarter of the number who contribute one time; the number who contribute three times is about one quarter of the number who contribute One-ninth of those who contribute, etc.; those who contribute n times are about one-ninth the number of those who contribute 1 time.”

Like Western textbooks, the Soviet informatics textbook adopted by Cuba provides an in-depth introduction to such “information laws”. Their lead authors, Russian information scientists and engineers AI Mikhailov and RS Gilyarevskii, cite American information scientist and historian of science Derek de Sola Pugh A passage from Derek de Solla Price on the distribution of the number of publications: "They follow the same type of distribution as millionaires and peasants in highly capitalist societies. A large part of the wealth is held by the very few extremely wealthy of individuals, and the remaining small part is in the hands of a large number of very few producers.”

This unequal distribution of information wealth must also be radically changed for Cuban information scientists who have gone through the socialist revolution and the sudden redistribution of material wealth. Among these information scientists is Emilio Setién Quesada, a scientist who has been researching and working with Freire de Andrade since the post-revolutionary period. Setion Quesada has challenged the idea of "information laws", in an article co-authored with a Mexican colleague, he objected to the term "law", which seemed to Means "to establish a causal, persistent, and objective relationship in nature, society, or thought." Mathematical equations represent "regularities" only, and do not express "the causes of the qualitative characteristics of the behavior they describe." These The reasons are historical, not natural.

Therefore, Setion Quesada and his colleagues argue that the number of publications does not ultimately determine an author's "productivity," any more than a decline in the number of citations indicates the "obsolete" of a publication. Cuban libraries should not rely on these criteria to make critical decisions about which materials to discard. Traditional informatics is incompatible with revolutionary library science because it treats the accidental regularities of history as immutable laws that tend to perpetuate existing social inequalities.

However, Cuban information scientists are not just criticizing the limitations of traditional informatics. They also propose a more critical approach to mathematical modeling that emphasizes the complexity of society and the historical contingency of the laws of information. In the 1980s, when Cuban libraries began to adopt digital computers, Setion Quesada was tasked with creating a mathematical model of library activity based on statistical data for economic planning. But he was dissatisfied with the models of the "intensity" and "effectiveness" of library activity devised by Soviet and American information scientists. (In the discussion below, I have put the mathematical explanation in parentheses for the interested reader, following Setion Quesada's own terminology and notation.)

Soviet information scientists calculated the library by multiplying the "coefficient of intensity" (the number of borrowings m divided by the number of potential readers N) by the "index of circulation" (the number of borrowings m divided by the total collection f). The "coefficient of intensity" of the activity. At the same time, American information scientists calculated the "measure of effectiveness" of the library, dividing the index of circulation and the "index of capture" (the actual number of library readers n divided by the potential The number of readers N) is combined. Comparing these two approaches, Setion Quesada proposes an alternative "Cuban model" that evaluates what he calls the "behavior of Cuban public libraries":

"coefficient of intensity"

"measure of effectiveness"

"Cuban model"

According to Setion Quesada, "the Cuban model is more complete." It includes more variables, all of which he believes are important. For example, the Cuban model included an "index of communication" (based on the number of readers l using the archive), while the Soviet and American models "did not express the precise degree of author-reader social communication that took place in the library". Furthermore, those other models "did not take into account the role of librarians in the development of activities". For Setion Quesada, librarians "form, together with readers, the main positive factor involved in the development of this activity". Therefore, in the 'Cuban model', each variable was adjusted relative to the number of librarians (included in the adjusted variables, indicated by line brackets). Finally, other models "did not provide an index of comparative behavior across regions and periods." In contrast, the "Cuban Model" attempts to facilitate comparisons across libraries and over time (each denoted by the subscript i).

Regardless of the strengths and limitations of this particular mathematical model, the broader story of Cuban information science encourages us to be skeptical of the claims attached to current information retrieval models and algorithms. If yesterday's information scientists claimed that their models ranked authors by "productivity" and libraries by "effectiveness," today's "artificial intelligence experts" claim their algorithms are based on " Relevance” (relevance) ranks “personalized” search results. These statements are never mere descriptions of the nature of things. Rather, these are interpretive, normative, politically significant stipulations about what information should be considered relevant or irrelevant.

Disguised as descriptions, these regulations reproduce an unjust status quo. Just as print publications should not be considered obsolete and discarded from library collections based on citations, online information should not be considered irrelevant and ranked low in search results based on "click-through rate" and advertising revenue. Innovative experiments by Cuban information scientists remind us that we can design other patterns and algorithms to disrupt, rather than perpetuate, patterns of inequality and oppression.

The Network Theory of Liberation Theology

The Cuban experiment was backed by a socialist state. But in the absence of such a state, experiments in anti-capitalist informatics are also possible. In fact, another major cause is taking place in a country controlled by a U.S.-backed right-wing military dictatorship.

In many Latin American countries, including Brazil after the 1964 military coup, authoritarian regimes resorted to violent measures to suppress dissidents, such as censorship, imprisonment, torture and exile. The strongest critics of these measures are Catholic priests, who seek to reposition the church to organize the oppressed and overcome domination. A key event in the formation of their movement, which came to be known as "liberation theology", was the Latin American Synod in Medellín, Colombia, in 1968. At the landmark conference, attendees learned about the dynamics of oppression in different countries and collectively declared, "deafening cries from the throats of millions, asking their priests to grant them emancipation everywhere."

How can this cry be heard? Medellín's experience inspired a group of liberation theologians, mostly from Brazil, to try to imagine new ways of communicating between poor and oppressed peoples around the world. Their goal is "conscientization": cultivating a critical awareness, including reflection and action, to change the fabric of society - a term related to their colleague Paulo Freire, who developed The theory and practice of critical pedagogy. To this end, the theologians plan to organize a series of conferences called International Journeys for a Society Overcoming Domination.

But international conferences are prohibitively expensive, meaning many are left out. Chico Whitaker, a Brazilian Catholic activist and one of the project’s organizers, explained that “international conferences are rarely free from domination: generally, they are reduced to ‘experts’ who are able to meet. Conference.” To address this problem, liberation theologians and coalition activists envision a system for the dissemination and circulation of information, which they call an “intercommunication network.” The network will provide "information that was not manipulated and without intermediaries", break down "sectoral, geographic and hierarchical barriers" and make "the discovery of controlled information systems deliberately undisclosed" a possible.

Through "controlled information systems," organizers point to the intense national censorship of print and broadcast media that prevails in Latin America. Liberation theologians want the liberation of information, which will bring Freire's pedagogy into a new phase: from the age of mediated "conscientization" to direct communication between the oppressed, in Whittaker's words The era of "inter-conscientization".

Since the modern Internet did not exist in the 1970s, the operation of the "intercommunication network" relied on print media and postal services. The organizers set up two offices, called "diffusion centers": one in Rio de Janeiro, at the headquarters of the National Synod of Bishops of Brazil, where the organizer of the Medellin conference, Bishop Cândido Padin of Brazil, serves as a project Coordinator; another in Paris, where Whitaker went into exile with his wife Stella, another Brazilian activist for his involvement in land reform planning before the 1964 military coup.

The "Communication Center" receives and distributes by mail short texts (or five-page summaries of longer texts) analysing the situation of "domination" from a global network of participating organizations through Latin America, North America, Africa, Europe, Asia Associated with the Regional Synod of Bishops in Oceania.

Whitaker stressed that these texts are best written by "those who are most interested in overcoming domination, that is, those who are governed," and should include "a sense of their own situation and their desire to liberate themselves from domination." Analysis of the struggle to come out". The organizers published each text that met the essential requirements without any editorial changes; translated each text into four languages (Portuguese, Spanish, French and English); and mailed all texts free of charge to over ninety participants in each country.

For Whittaker, the concept of mutual communication is rooted not only in the "freedom of expression" but also in the "liberty of information": all participants have the communication, which contributes to the achievement of their common goals.” Mutual communication seeks to produce complete equality: “All must be able to speak and be heard, regardless of each person’s rank, level of education or experience, social function or status, morality, knowledge, or political authority." Whitaker writes that the practice of mutual communication requires "acceptance of heterogeneity and the 'dynamic' conflict that accompanies it."

Finally, mutual communication requires "mutual respect" and "openness to others," which reflects the Christian principles of fraternity: as Whitaker puts it, "respect the thoughts or actions of others...accept the new and the unexpected, accept Something that asks us questions or challenges, or accepts views and concerns that we could have dismissed because they are hard to accept.” While Christian values are important, the “intercommunication network” is open to anyone of. Some of the participants were non-Catholic, non-Christian, or even non-religious. "As God's children, we are all brothers in Christ without any distinction," Patin explained.

freedom to be heard

Over the years, the "intercommunication network" has spread various texts. Participants in Chad studied the social consequences of cotton monoculture since the introduction of cotton monoculture under French colonial rule. Participants in Sri Lanka reviewed labor conditions in the fishing industry, profiteering strategies of seafood exporters, and the limitations of state-established fishing cooperatives. Participants in Panama recounted their struggles for housing and the neighborhood associations they formed. In Guinea-Bissau, a group of local and foreign educators, including Paulo Freire, wrote that they faced challenges in organizing literacy programs and changing the education system after the war of independence. Between 1977 and 1978 alone, nearly a hundred articles circulated on the Internet. These were later compiled into a tome, published in four languages, and discussed at regional meetings of network participants around the world.

This book employs an unusually complex indexing system. Each text has a code consisting of letters and numbers; for example, the code for the aforementioned Chad text is "e35". Letters indicate the type of text - "e" for case study, "d" for discussion text, "r" for abstract - and numbers in chronological order. The book is divided into sixteen numbered chapters, each on a different topic of "domination". The third section focuses on "control over rural workers," the fourth on "non-rural workers," the seventh on "domination of housing conditions," and the tenth on "sanitary conditions."

Each text is printed within one of the subject sections, but since the classifications are not mutually exclusive, the index of each section also lists texts that intersect the subject, albeit from different sections. For example, the index to Section IX on Education lists some of the main texts—“e4” in Thailand, “e6” in Guinea-Bissau, “e38” in the Philippines—as well as other texts from different chapters, such as the tenth "r3" of the section, which discusses the intersection of health and education in governance structures. An index was added at the end of the book, classifying texts according to "certain categories of victims of domination": "women," "youth," "children," "older people," and "ethnic groups."

The astonishing diversity of texts disseminated by the "intercommunication network" quickly brought its organizers into conflict with the conservatives of the Catholic Church. In 1977, some readers were particularly outraged by "e10," a small female-led group that described itself as a "community of Christian love" in rural England. The article upset conservatives not only because it explicitly condemned "the Roman Catholic Church as an instrument of domination" engaged in "an effective and specialized form of 'brainwashing'", but also because of its feminist proposals, including rejecting " Calling anyone 'Father' in the clergy", as well as committing to "calling the Holy Spirit 'She' and not 'He'".

After lengthy discussions at the "diffusion centers" in Rio de Janeiro, the project organizers decided to publish the text with a note reaffirming their commitment to freedom of expression and reminding readers of the minimum requirements for publication. Still, conservative bishops have complained to Vatican authorities, which are increasingly concerned about the rise of liberation theology in Latin America and elsewhere. Pope Paul VI was not sympathetic to the project, sending envoys to Brazil to intervene. The Vatican asked the bishops to stop, claiming that the Rio de Janeiro conference "cannot take such a broad initiative and that it is beyond its purview to invite other bishops' conferences to join the project." Building a distributed global network through regional conferences, liberation theologians Bypassing the central authority of the Vatican. Despite the Vatican's order to stop the project, a group of Brazilian organizers continued the fight until 1981.

Later, the former organizers reflected on the relationship between their "intercommunication network" and the modern Internet. Little did they know, in the original paper on the Transmission Control Protocol (TCP), outlining the technology that underlies the Internet, engineers Vinton G. Cerf and Robert E. Kahn talked about A protocol for grouping "network intercommunication," or simply "internetwork," from which the word "internet" came a few months later. The paper was published in 1974, at a time when liberation theologians were planning their own similar network.

In 1993, when Chico Whitaker reflected on the two Internets, he proposed that "network" is an "alternative structure of Less common than the "pyramidal structure":

Information is power. In a pyramid structure, power is centralized, as is information, which is hidden or kept for use at the right moment, thereby accumulating and concentrating more power. In the web, power is decentralized, and so is information, which is distributed and disseminated so that everyone has access to the power they have.

There is no doubt that Whitaker and his colleagues lean towards techno-utopianism. They hope that technological progress will eventually enable the "free" flow of information, which is an illusion, because the various machine decisions and human labor constituted by political and economic conditions always filter the content and objects of information flow. The techno-utopian conception of "information freedom," neither in California's version of libertarian-capitalist nor in Brazil's version of liberation theology, is never quite correct.

However, there is a key difference between the two concepts. The California version of freedom of information is largely limited to a specific understanding of freedom of speech. Silicon Valley companies that govern the public discourse on the internet, such as Facebook, insist on "free speech" as an excuse to profit from posts and ads that spread right-wing misinformation.

The remarkable innovation of the Brazilian liberation theologians is that they moved beyond a narrow focus on freedom of speech to a politics of the audible. Theologians understand that the question is not just whether one can speak freely, but who one can hear and which audiences one can reach. The purpose of the "intercommunication network" is to create fairer conditions not only for speaking, but also for listening and being heard. Ultimately, the purpose of the network is to amplify the voices of the oppressed. Today, our task is to reformulate this more critical concept of freedom of information for the digital age. Information is "free" only when the voice of the oppressed is as loud as the voice of the oppressor.

The return of history

The history of technology is often described as a linear progression, a series of stories of successful inventors, mainly from North America and Western Europe. Stories like this are ubiquitous, in part because they are easy to tell. After a technology becomes popular, storytellers can simply follow the records and narratives given by a few people who have already been credited for their inventions.

This trite narrative has an important ideological function. First, they legitimized capitalist accumulation by defining the inventor-entrepreneur's wealth as the due reward for an idea. This requires erasing all other contributors to a particular technological product; in the case of search engines, that means forgetting about librarians (whose feminized labor was never considered creative) and information scientists, who for decades 's cumulative work lays the foundation for Google.

Even more subtly, this narrative also endorses mainstream technologies because they are the only ones imaginable. They ignore many possible alternatives that haven't caught on, creating the impression that existing technology is simply the inevitable consequence of technical ingenuity and good judgment.

If fringe innovations like the Latin American experiment in informatics do not go mainstream, it is not because they are necessarily inferior to their corporate, military and metropolitan competitors. The reason why some technologies survive and others die is not strictly technical, but political. The Cuban model is arguably more complex than the American model. However, some technologies are sponsored by the advertising industry, while others are restricted by neocolonial trade embargoes. Some are backed by the Pentagon, while others are suppressed by the Vatican.

Retrieving those lost replacements is crucial because they show us that technology could have been and still can be in the future. However, these histories are difficult to retrieve. Their protagonists may remain anonymous, and their records will not be made public.

No search engine pointed me to the Latin American experiment. With traditional internet search methods, I would never have been able to find them. Instead, I found some subtle clues through casual conversations. I chat with Theresa Tobin, a retired MIT librarian who co-founded the Feminist Task Force at the American Library Association in 1970. She commented that after she raised funds in the 1980s to donate a digital computer to the Sandinista library, Nicaraguan librarians used it to implement Cuba's system of indexing materials.

I started to learn more about the Cuban system and it was a tough job. The most important sources of information on Cuban information science are difficult to find using traditional search engines and databases. For example, despite María Teresa Freyre de Andrade's fame, Google Scholar does not index her main books, and Wikipedia does not have any of her languages entry. On the other hand, the Cuban online encyclopedia EcuRed, has an article about her. I also managed to find some initial references on Cuban informatics in the Latin American bibliographic database SciELO. I then contacted Cuban academics directly for help.

My discovery of the "intercommunication network" of liberation theologians follows a similar path. When I first met Stella and Chico Whitaker at the World Social Forum in Porto Alegre, I never heard Said the "intercommunication network", the forum they co-founded in 2001. It wasn't until many years later, when I helped the couple donate their personal documents to the public archives, that they mentioned in passing that a dusty box in their apartment contained documents from an old project involving informatics. They were surprised that I showed interest. Sometimes the best way to retrieve information is to talk to people.

More big ideas for alternative futures, technical or otherwise, are still left in dusty boxes around the world. The pent-up dreams of past struggles do not easily appear in the algorithmic feeds of our businesses. To retrieve these lost ideas, we must develop more critical information retrieval methods, continuing the unfinished work of the Latin American experiment. In short, we need critical search.

The critical search project will recognize that any quantification of "relevance" is an interpretive, normative, and politically relevant act. Critical Search will actively work to increase the visibility of anti-hegemonic intellectual traditions and historically marginalized perspectives. We must build systems of information dissemination and circulation that seek to amplify critical voices across language, ethnic, racial, gender, and class barriers. Let us draw inspiration from our predecessors and strive to follow in their footsteps. Let's reindex the world with algorithms, interfaces, and strategies.