TITRE : The symbol grounding problem 75 years after Turing's Test (why computational success still leaves meaning unexplained)
RÉSUMÉ
This talk examines the enduring symbol grounding problem 75 years after Turing's seminal 1950 paper, questioning why computational success in language tasks fails to resolve fundamental questions about meaning. Strohmaier will explore how large language models' impressive linguistic capabilities paradoxically highlight rather than solve the challenge of connecting computational representations to real-world meanings. Drawing from recent work in computational lexical semantics and philosophical analysis of machine understanding, the discussion will address why statistical pattern learning, despite its empirical success, leaves core questions about semantic grounding unanswered. The talk bridges computational linguistics and philosophy of mind, examining whether computational approaches can ever truly capture the referential and intentional aspects of human meaning-making, or whether meaning remains fundamentally unexplained by computational success alone.
BIOGRAPHIE
David STROHMAIER is a Research Associate in the Natural Language and Information Processing (NLIP) group at the University of Cambridge, supported by the Institute for Automated Language Teaching and Assessment (ALTA). His research applies machine learning and deep learning to lexical semantic acquisition, investigating how neural models learn word meanings and their relationship to human learning processes. Strohmaier has a highly interdisciplinary background, holding a PhD in Philosophy from the University of Sheffield (2018) and an MPhil in Advanced Computer Science from Cambridge. His work bridges computational linguistics and philosophy, with publications spanning word sense disambiguation, computational semantics, social ontology, and decision theory. He is co-author of "Preference Change" (Cambridge University Press, 2024).
RÉFÉRENCES:
Strohmaier, D., & Messerli, M. (2024). Preference Change. Cambridge University Press.
Strohmaier, D., & Wimmer, S. (2023). Contrafactives and Learnability: An Experiment with Propositional Constants. In Post-Proceedings of Logic and Engineering of Natural Language Semantics 19, 67-82.
Strohmaier, D., & Tyen, G. (2022). A Category Theory Framework for Sense Systems. In GLOBALEX 2022 @ LREC.
Strohmaier, D. (2021). Organisations as Computing Systems. Journal of Social Ontology, 7(1), 1-25.