The lecture by Dr Michael Ramscar from University of Tübingen will take place at the Faculty of Arts, Charles University (Na Příkopě 584/29, Praha 1, room no. C502) on 17 June 2026 at 10:00 am–12:00 pm.

Abstract

Traditional studies of language assume an atomistic model in which linguistic signals comprise discrete, minimal form elements “associated” with discrete, minimal elements of meaning. Production has been seen to involve the composition of messages from an inventory of these form elements, while comprehension is the subsequent decomposition of these messages. Researchers have thus focused on attempting to identify and classify these elements, using words like “encode,” “semantics” and “communicate.” These words are used in the absence of clear formal definitions, the assumption being that these will emerge out of further study.

By contrast, information theory has provided precise mathematical definitions for at least some of these terms, albeit within a formal framework that seeks to minimize uncertainty as a message sent in a signal is discriminated from the possible messages that might have been sent. In a similar vein, behavioral and neuroscience research has revealed that the basic processes that enable animals to learn “associations” are also discriminative, producing models that serve to minimize an organism’s predictive uncertainty about the environment.

In this talk, Dr Michael Ramscar will describe some empirical results that indicate that human communication is subject to the constraints that the basic principles of information and learning impose, and describe how, from this perspective, languages should be seen as discriminative communication systems. In illustrating how this process works, I will show how a discriminative approach to communication makes sense of many aspects of language that have long seemed puzzling, such as noun class systems (colloquially known as grammatical gender) and the semantics of personal names. I will show how noun class and personal name systems actually represent highly structured and highly evolved linguistic subsystems that offer remarkable insights into the discriminative nature of human communication.

Photo: ProviderNation