Giulia Rambelli to give Semiotic Syntax talk on analogy in language processing W11/19
Live from Bologna...it's Dr. Giulia Rambelli, discussing the role of analogy in language processing. Come to the watch party in person or attend via Zoom on Wedesday, November 19, 10-11:30am in the Syn-Sem Lab (LBB 151). Attend in person and enjoy refreshments and the fellowship of the Syn-Sem Lab or attend remotely via Zoom: . Further information about Dr. Rambelli's talk can be found below.
Analogy in Language Processing: A Constructionist and Computational Approach to听Productivity
Abstract
Understanding the mechanisms that people use to go beyond their known input to produce novel, structured utterances鈥搕hat is, to achieve linguistic productivity, is a crucial domain in linguistics. This talk will explore how novel meaning is constructed by relying on patterns of use rather than by fixed compositional rules. Within a constructionist framework, analogy听is seen as the cognitive process that allows speakers to extend existing constructions to novel contexts, balancing creativity and conventionality. I will discuss insights from my recent work on the interpretation of novel noun鈥搉oun compounds听in both humans and large language models, shedding light on the parallels and divergences in their processing strategies. The results will serve as a basis for a broader discussion of analogy as a cognitive mechanism underlying linguistic productivity, addressing two crucial challenges: 1) define specific tasks to investigate whether analogy plays a role in human comprehension, and 2) use Language Models as tools to investigate how the exposure to item-specific structures guide the interpretation of novel linguistic ones.
Bio
Dr. Giulia Rambelli听holds a joint PhD in Computational Linguistics听from the University of Pisa (Italy)听补苍诲 Aix-Marseille University (France). She has been a Postdoctoral Researcher on the European ERC project ABSTRACTIONat the University of Bologna听补苍诲 will join the Institute of Language, Communication and the Brain (Aix-Marseille University) in January to carry on her own project 鈥淎NALOGIZE- Modeling Analogical Generalization in Children鈥檚 Creative Compound Formation through Computational Models.鈥 Her research sits at the intersection of theoretical linguistics, cognitive science, and computational models of language, with a focus on the mechanisms underlying sentence interpretation听补苍诲 linguistic productivity听from a usage-based perspective. In her Cambridge Element 鈥淐onstructions and Compositionality: Cognitive and Computational Explorations鈥, she examines the balance between compositional and direct access to meaning, exploring how the cognitive process of analogy听can be modeled as a source of linguistic productivity.
听
