By Louis DiPietro

Growing up in a Dalit community in India, Khushi – a software engineer in her early 30s – was taught to keep her caste under wraps.

But one day at her job, a project manager insinuated in an annoyed tone that a Dalit woman had unfairly received a government job over his wife because of benefits or reservations afforded to those in lower castes. A Dalit herself, Khushi confronted the manager, even as speaking up meant unveiling her position within India’s caste system and risking her future at the company.

Stories like Khushi’s are common among Dalits and lower-caste Indian people working in computing around the globe. According to award-winning research from Cornell scholars in information science, caste remains an inescapable construct requiring careful navigation, interpretation, and interruption.

In sharing stories from Dalits, the Cornell team shines a light on caste marginalization in computing and human-computer interaction (HCI), with the intent to further discussions of equality in the field and raise awareness of the ways caste emerges in the everyday culture of computing and HCI, as well as in the design of technological systems. 

“Caste affects how South Asians think, talk, and relate to other people. It’s a state of being and knowing the world,” said Palashi Vaghela, a doctoral student in the field of information science and lead author of “Interrupting Merit, Subverting Legibility: Navigating Caste in ‘Casteless’ Worlds of Computing.”

Coauthored by Steven Jackson, associate professor of information science, and Phoebe Sengers, professor of information science, the research received a Best Paper Award at the ACM CHI Conference on Human Factors in Computing Systems, held in late April. In a separate paper to be presented at the ACM Conference on Computer-Supported Cooperative Work and Social Computing in November, Vaghela explores how caste manifests as social capital on Twitter among politicians in India.

“How do we make sense of something that people assume is outdated but still determining our everyday lives?” Vaghela said.

In the CHI paper, Vaghela and coauthors draw on interviews with 16 Dalit engineers working in India, the U.S., and the U.K., including Khushi, the software engineer who interrupted arguments of merit and caste by her manager. Each tells of the subtle and overt ways in which caste surfaces emerges for Dalit and other lower-caste Indians during schooling and later in the computing industry – like navigating sensitive conversations about religious beliefs that could signal caste positioning, considering company meal options that are more customary for those in upper castes, and fielding direct questions about rank in engineering exams.

“I could have complained to the HR about this, [and] there would have been action” said another Dalit engineer and interviewee, recalling how a coworker suggested he attended “anti-national festivals,” an implication that he was anti-caste and thus averse to the idea of Indian unity. “But all my friends would, like, stop talking to me if something drastic would have happened to him. There's peer pressure … you can’t go against people who are deliberately making these comments.”

Their stories reveal how Dalits navigate and sometimes obfuscate their caste from Indian classmates and coworkers, most of whom come from upper castes, and connect with other lower-caste people in their work organizations.

“The more you climb up the professional ladder, especially in computing and engineering, the number of people from similar backgrounds is fewer,” Vaghela said.

The authors offer practical ways of addressing caste to build a more open and inclusive culture in computing:  

  • Understand the sensitive nature of conversations involving caste among team members, knowing that caste continues to shape the lives of Indian practitioners in HCI and computing. 
  • Explore better industry data on caste and religion composition of computing practitioners. This would help inform a “caste-aware” look at the fabric of caste and its norms within the field, the authors wrote. 
  • Survey the perception of affirmative action policies, like reservations for lower-caste communities, in the computing industry.  
  • Support efforts that allow lower-caste engineers to safely find and build community networks within their organizations on their own terms. 
  • Understand that the caste hierarchy structure is complicated by many socio-cultural and economic factors, particularly colonial categorization that has a bureaucratic life of its own.  
  • Reframe the issues of caste in computing as a privilege that benefits those in upper castes, and redirect the responsibility of transformation onto those who benefit from caste’s institutionalization.

“What we emphasize is that identities, and thus by extension people who face historical discrimination, are not static data points,” Vaghela said. “There has been a tendency to frame Dalit stories under the idea of damage, shame, and violence in scholarship; our study shifts the singular focus on the notion of damage to artfulness and agency of Dalits in a way that makes room for a different interpretation one of resilience and subversion that plays with the legibility of caste in the worlds of computing.”

This research is supported in part by the Social Science Research Council and the National Science Foundation.

Louis DiPietro is a writer in the Cornell Ann S. Bowers College of Computing and Information Science.