The iSchool Research Symposium is back. After a pause, the series returns this fall with talks most Fridays, bringing together researchers from across the iSchool and beyond to share new work, spark conversation, and connect our community with the wider world of information research.

We're opening with Allen Renear, Professor in the School of Information Sciences at the University of Illinois Urbana-Champaign, where he served as dean from 2012 to 2019. Across a career spanning text encoding, data curation, formal ontology, and the philosophy of computer science, Renear has asked what our information systems actually mean, and what they quietly assume. His talk takes on one of the biggest stories in the field, and asks whether large language models have resolved its oldest problems or just hidden them.

Abstraction, Paradox, the End of History … and the Re-emergence of the Vernacular

Abstraction is the unifying ideology for the triumphalist history of the progress and success of the information sciences generally, and computer science in particular. It plays the leading role in how we tell the story of the evolution of programming languages, software engineering, data management systems, and document text encoding. More generally, it exemplifies the search for generality, efficiency, and elegance that characterizes the information sciences. During the 1990s and early 2000s, it appeared that this progress was approaching its culmination, promising a perfect match between human understanding and the design of digital information systems. But familiar paradoxes of abstraction began to reappear, and what was once seemingly idle philosophical amusement in ancient Athens, Nalanda, Baghdad, and Fez now suggested serious threats in modern automated digital environments. There appeared to be no obvious general resolution, only mitigation through ad hoc human intervention. Today, a new and unexpected chapter in this history has begun: the emergence of large language models. Instead of the hoped-for impedance match at the level of explicit, if brittle, abstract formalizations, person and machine carry on their discourse in the vague, suggestive vernacular, each exploiting nuances the other need not explicitly articulate and each benefitting from indeterminacies the other need not formally resolve. Or have the fundamental issues merely been hidden – and perhaps made more dangerous – by an unaccountable simulation of understanding? 

Refreshments provided. Come early, grab a snack, and settle in. Paradoxes are better with a little something to nibble on. (Arrive early for best selection.)