Beneath the fingers of the petites mains

Slides #

ouvrir dans mon navigateur

Text #

In this presentation, I would like to share with you one of the outcomes of this research, a research to be continued that is focused on what I called the “petites mains” and the Human/Non-human distinction.

I kept the French expression “petites mains”, because it is quite complex to translate as it is in English: small or little hands would not convey the meaning of the French formula.

(Maybe the term amanuensis, or textual assistant will be closer, but I’m afraid it’s mainly a matter of textual environments, although les petites mains reach beyond that.)


The “petites mains” referred in the 19th century to a labor grade in sewing industry :

In clothing workshops, the first hands, les premières mains, assigned meticulous and repetitive tasks to the petites mains that were apprentice sewer, exclusively women apprentice.

Today we use this common expression to qualify all activities that are hand-made, tasks that require as much time as energy. The “petites mains” are the anonymous actors in the backstage of production, theirs hands are precise, efficient, almost invisible, and so discreet that they don’t matter to the eyes of science, it’s part of the machine, of the non-human dimension.

For the last ten years or so, there has been a peculiar concern in studies around the behind-the-scenes of science, and the term “petite main” has become as much a way to express a revalorization work as a statement. Among all the studies that can be valuable sources to my project, I cannot quote all of them, and if I don’t cite you, don’t blame me, I’ll be happy to read your research and very open to other reading suggestions for a work that is much still in progress.


More than any other fields of production, Humanities perceive knowledge as a human matter, but, with the academic epistemological model, they ignore the conditions of emergence of their own thinking, considering the participation of the petites mains as irrelevant to the establishment of knowledge and devoid of scientific expertise.

Thinking, producing ideas is not, does not emerge from doing, from factories, in the classical knowledge production model. There can be only one creator, and not a group of apprentices and secretaries.

(We know it since Plato, ideas have a world of their own, and comes basically from cloud).

If the digital humanities, the latest humanities, were an opportunity to argue and redefine a scientific model, the pattern of knowledge remains quite the same.

How can we renegociate with a tradition of illusion and shadows ?


Despite the Digital Humanities promises of knowledge production tools and processes inclusion, the figure of the lone scholar is still strongly present within its origins.

Mea culpa for citing a Digital Humanities project that is quoted too much compared to other projects that are not quoted at all. I chose not to take another project as an example because the Index Thomisticus project benefits from being extensively studied, analysed in its archives and also de-constructed, thanks in particular to the work of Melissa Terras, Julianne Nyhan and Steven Jones.

The Index Thomisticus project was a Father Busas project to put into a form that was machine-readable the work of Thomas Aquinas. It’s not just an exemple of collaboration between a man of science and a powerful company that was IBM, or a collaboration between religion and technology, or just a promising gathering of whites men, it’s although a great example of a scientific pattern.

Most fields cannot point to a single progenitor, much less a divine one, but humanities computing has Father Busa.

(phew)


It is neither the IBM men nor Father Busa and his faith that have concretely set the works of Thomas Aquinas into punchcards.

In the shadow of the great man and the powerful company, it was in fact 65 young female operators who did the work of encoding.

These ghost workers - whose names do not appear in the archives - actually made the project happen: their work enabled the model designed by Busa to be renegotiated, expanded and challenged according to technical realities.


That being said, does the fact that female workers encoded the punchcards means that the project is no longer Father Busa’s?

While the punched card operators relied on the input of the scholars to guide the transposition of the text from container to container, so the scholars relied on the keypunch operators to reify the Index Thomisticus and related texts as scalable, computable artefacts. One could not properly do their work without the other; both played fundamental roles in the process of data capture and dataset elaboration.

It’s not a matter of removing Busa from the project, it is a concern to have better understanding of a model that the Index Thomisticus embodies for the culture of the digital humanities.

The myth of the lone scholar, which Busa himself helped to build, disguises the collaborative dimension of the project. But this collaboration was not without a sexist hierarchy.

Nyhan interviews with former keypunch operators that could be identified 2014 shows that women represented at the time a low-cost and low-skilled ressource of work. They were left without the opportunity to move up to a higher position than punchcard operator, and most of them had not even been informed of the final aim of the project.


About this project, What surprises me even more is that the Busa archive, that collects a large part of the documentation on the project, began in 2009 under Father Busa’s supervision, and was continued after his death in 2011, and so these ghosts of the project were there : the women on the photographs are neither out of focus nor hidden. The fact that they are not seen is an optical and epistemological problem. We don’t see them because the “petites mains” have, in the dominant cultural model, no visibility, that is no ability to be seen.


The patriarchal logic of the myth continues in many projects kept in mind by the Digital humanities and validates the importance of the document (of scientific research more than of creative exploration as pointed out by McPherson in Feminist in a Software Lab).

Here an example of another important myth of man epistemology and master-narration, the Mundaneum, the project of indexing the world literature to instore peace betweens humans, and the image speaks for itself : there is A or even The Man of the project, Paul Otlet, and then there is the petites mains.

It’s about document and even more about documen.


The reality of technical sexism, which is the subject on which Isabelle Collet has worked extensively, is that women are, since these fields are no longer part of the service sector, mainly excluded from technical training and environments, even from technical imaginations that are essentially masculine (geek, programmer, hacker).

The introduction of microcomputers has produced hackers’ societies […] almost entirely men’s and anti-girl. […] The hacker has become the ideal model of the computer scientist. Not only does this career no longer match with the image that girls have of themselves, but it even appears to be frankly hostile to them.

This conception shapes our perception: when we see a woman behind a screen, she is not coding in python or in haskell or even in html, she is treating emails for a higher ranking man. She is not a geek, a programmer nor a hacker, she is a secretary or a human computer.


There is this persisting conception in our studies that strikes me more and more, that is the working women figure is a construction. The working woman exists to be at the service of others, she justifies its existence as an extension, a technical extension of or for men, just as media in McLuhan words.

We find this idea in the Viliers de l’Isle-Adam Novel The Future Eve or Tomorrow’s Eve where a fictionalized Edison build the perfect woman as the first women robot for a Lord desperatly in love.

(The real woman that the Lord loves is beautiful but silly, so Edison will build the same woman, in appearence, but with a brain, or artificial intelligence, trained on literature.)

The novel describes all the mechanics, the technical implementation of what a woman is or what she should be in the eyes of men, even explaining how to implement the female pleasure on a technical level, because it seems to hard to find and get in real life.

This myth of the andreid is not just fictional, we find it also in the designs of the first chat agents that turn out to be 2.0 secretaries, extensions of man-woman domination relationships, that is obedient, polite, submissive, flattered when insulted (the “I blush if I could syndrom” witch was for more than 10 years the answer of Siri when insulted).

On this topic, I highly encourage you to follow Lai-Tze Fan and Hasti Atapour’s work on The Evolution of Siri’s Sexism and Apple’s Corporate Social Responsibility.


The place where I think punchcards operators, computers, and andreides meet, is a pattern based on service, but also on a hierarchy between thinking, academia, human and making, factories, non-human.

It is precisely on this aspect that I would like to contribute but, it can sound as a paradox, my approach does not aim directly the petites mains revalorisation.

First of all, I cannot pretend to revalue the participation of human computers, keypunch operators or even today’s working women in computing as one single entity : as Valérie Shafer reminds us in Connecting Women, social and political contexts are obviously very different in each case.

We cannot get justice by blending status.

However, there are, I believe, some common issues that come from science patterns: all women in these fields are paid less than men in the same profession, and there is a broad principle of devaluation of their work. Issues of their non-crediting and “invisibilization” are not the core of the problem. They are only symptoms of a deeper concern: the conception of knowledge emergence implying that thinking comes first and is an abstraction.


I identify several potential drifts in the revaluation approach: first, the risk, as highlighted by Crystal Bennes in Klara and the Bomb, of producing an alternative history, and of creating a caste of exceptions.

Like Sadie Plant, in her book Ones + Zero, Bennes’ argument is that the production of a feminist history that replaces male figures with female ones will not impact the problem of the dominant history at the roots.

There is also the risk of creating mythical figures as imposing as Father Busa.

We don’t need mothers, or authority figures of lone scholar, we need communities models.

Besides, not all women who worked in the development of computer science were Ada Lovelace or Grace Murray Hooper, and not all women were hidden by a system, some agreed with the system in place.

Above all, there is the risk of glorification, maybe the most insidious one: among the computers and operators, some of these women have contributed to destructive activities by providing data to test various bombs (like the ENIAC six).

[…] just because women were involved in this work, and that their work has been suppressed, it does not mean they should be celebrated as feminist heroes. The work these women did had serious and damaging consequences in the real world, and I cannot talk about them without talking about the appalling impact of their work.


On a more concrete level, listing all the actors in the production of an article, would have the effect of what I would call a film credits: in cinemas, no one stays to look at the names of all the people, and only the name of the director as the unique and lone producer is kept in mind.

To say with McPherson, to add is not to include and since there is an epistemological model at the foundation of power and domination dynamics, it is unlikely that the issue will be resolved in a single generation time. (sadly but we will keep working on it)

Many of the archival recovery efforts in the early years of DH deployed a similar additive logic, despite their good intentions. When these efforts focused on adding race or gender to digital archives and data sets, there was an implication that simply adding new data as content is all that is needed to get at some truth about race or gender. While it is hard to argue against, for example, including women authors in a database of nineteenth-century writers, such an approach is more additive than integrative or relationnal.

The issue is therefore not so much to reshape myths and represententions, as to challenge patterns and deconstruct/debunk a model that conceals presence and entanglements between human and non-human.


This lead me to conclude by opening the petites mains beyond the feminist reading that I have kind of perform.

The invisible work issue raises the problem of the single definition of the human categorie, a categorie defined by the difference with what is not human, what is non-human.

In other terms :

It’s not a question of wondering whether ChatGPT or other language model thinks and writes better than a human, or of deciding which of human or artificial intelligence is superior, it’s a matter of understanding that behind every data model, there are human patterns of exclusion and inequalities.

Depending of the boundaries uses to define the human categorie, we divide a community of knowledge who in fact is collaborating and blind us from the concrete dimension of our thoughts.


CC BY-NC-SA Antoine Fauchié