The Balloon

DIFFERENCE ENGINE

Labor and mechanized calculation

Breathless media coverage of machine learning tools and their applications often obscures the processes that allow them to function. Time and again, services billed or understood by users as automatic are revealed to rely on undervalued, deskilled human labor.

There is rich historical precedent for the presence of these “ghosts in the machine.” In a 2017 lecture, Director Emirata of the Max Planck Institute for the History of Science LORRAINE DASTON examines the emergence of mechanical calculation, revealing a fascinating history of the interaction between new technologies and the methods of routinizing and dividing intellectual labor that emerges alongside them.

From the introduction:

“The intertwined histories of the division of labor and mechanical intelligence neither began nor ended with this famous three-act story from pins to computers via logarithms. Long before Prony thought of applying Adam Smith’s political economy to monumental calculation projects, astronomical observatories and nautical almanacs were confronted with mountains of computations that they accomplished by the ingenious organization of work and workers. What mechanization did change was the organization of Big Calculation: integrating humans and machines dictated different algorithms, different skills, different personnel, and above all different divisions of labor. These changes in turn shaped new forms of intelligence at the interface between humans and machines.”

Link to the paper version of the lecture. (And stay tuned to the Phenomenal World for our upcoming interview with Daston.)

• A 1994 paper by Daston entitled “Enlightenment Calculations” gives specific attention to the logarithmic tables of Gaspard de Prony, which sought to demonstrate the usefulness of the newly-invented metric system: “The tables marked an epoch in the history of calculation but also one in the history of intelligence and work.” Link.
• Matthew L. Jones, an historian at Columbia University, studies the history of calculation and computing. His 2016 book Reckoning with Matter: Calculating Machines, Innovation, and Thinking about Thinking from Pascal to Babbage traces the history of attempts to routinize, mechanize and apply the power of calculation. Link to the book, link to Lorraine Daston’s review in Critical Inquiry.
• Simon Schaffer’s 1996 paper on the relationship between Charles Babbage’s calculating engine and the contemporaneously emerging factory system. Link.
• A syllabus prepared by Mary L. Gray and Siddharth Suri, authors of Ghost Work—a book about the “hidden” labor force behind many tech services—surveys the tech platform subcontracting labor market. Link.

New Researchers

Collective bargaining’s effects in Brazil

In his job market paper, PhD candidate in economics at Columbia LORENZO LAGOS studies the impacts of collective bargaining in Brazil, with a focus on an understudied factor: amenities.

From the abstract:

“This paper studies how a collective bargaining institution affects workers’ wages and amenities. The particular institution of interest is automatic extensions, which dictate that negotiated benefits cannot be phased out except through new collective bargaining agreements (CBAs). The introduction of automatic extensions in Brazil shocked the collective bargaining process by 1) allowing unions to holdout during negotiations; 2) expanding the time horizons over what is negotiated; and 3) increasing the relative value of amenities to wages. I find that mean wages fall at establishments with extended CBAs… and that bargaining units with strong unions secure additional amenities valued above the decrease in wages.”

Link to the abstract, and to Lagos’s website.

Each week we highlight great work from a graduate student, postdoc, or early-career professor. Have you read any excellent research recently that you’d like to see shared here? Send it our way: editorial@jainfamilyinstitute.org.

+++

• Souraya Sbeih and William Cook report on new data revealing how 400 Kenyans use mobile banking service M-Pesa. Preliminary findings: M-Pesa is quickly and frequently converted to cash. Link to an overview of the data (and link also to our previous newsletter on fin-tech and poverty alleviation).
• CEPR’s interactive “Blue Collar Jobs Tracker” shows the geographic decline in industrial jobs by sector and month, since 2018. Link.
• An engaging study of the iPhone from Tricontinental: “One estimate suggests that if the iPhone were made in the United States, it would cost at least $30,000 per phone.” Link. ht Francis
• Marta Kołczyńska empirically tests the influence of inequality on protest participation. Her conclusions: “Most of the observed effect of economic inequality on participation is due to between‐country differences rather than within‐country changes in inequality. The results therefore do not support a causal relationship between protest participation and economic inequality.” Link.
• “I am more impressed by people who make modest progress on questions of obvious importance than people who make decisive progress on questions that aren’t even worth asking.” Political scientist and anthropologist James Scott on the limits to scientific reasoning in the social sciences. Link.
• Regardless of whether they’ve taken a statistics course, college graduates fail to draw on prior research when calculating the probability of a reported research outcome. By W Burt Thompson, Amanda Garry, John Taylor, and Milen L. Radell. Link.
• From Columbia’s Center on Sustainable Investment, a new report outlines how value-chain shifts and climate risks are impacting the lives of coffee producers. Link.
• Maria Claudia Soler analyzes the experience of ISA recipients across Latin America: “Students who took ISAs expect higher future income, but have a worse academic performance than non-takers, suggesting that adverse selection might be a challenge for the large-scale implementation of ISAs.” Link.
• At StatsBlog, a critique of the existing definition of P-values offered by the National Academy of Sciences, which fails to “distinguish likelihood from probability.” Link. ht Sidhya
• Dylan Matthews skillfully breaks down the ramifications of Henrik Kleven’s new paper on the labor supply effects of work requirements. Link to the article, and link again to the paper, which we covered in a previous newsletter.
• “By focusing on the history of white ants in colonial South Asia, this article shows how insects were ubiquitous and fundamental to the shaping of British colonial power. British rule in India was vulnerable to white ants because these insects consumed paper and wood, the key material foundations of the colonial state. The white ant problem also made the colonial state more resilient and intrusive; government intervention was extended in order to control it. But despite state intervention, white ants did not vanish altogether, and colonized and post-colonial South Asians used white ants to articulate their own distinct political agendas. Over time, white ants featured variously as metaphors for Islamic decadence, British colonial exploitation, communism, democratic socialism, and, more recently, the Indian National Congress. Co-constitutive encounters between the worlds of insects and politics have therefore been an intrinsic feature of British colonialism and its legacies in South Asia.” White Ants, Empire, and Entomo-Politics in South Asia, by Rohan Deb Roy. Link.

Each week we highlight research from a graduate student, postdoc, or early-career professor. Send us recommendations: editorial@jainfamilyinstitute.org