LLF – Bât. ODG – 5e étage – Salle du conseil (533)
Mora Maldonado (Edinburgh)
Semantic Universals : Processing and Learnability
How do languages express meaning? Answering this question requires not only identifying the possible meaning(s) of every sentence and the semantic mechanisms that generate them, but also assessing which of these semantic mechanisms are shared cross-linguistically and which ones are not. In this talk, I propose to tackle these questions by adopting what I call a processing/learning approach: that is, I propose to investigate how these meanings and mechanisms are accessed, computed and learned by speakers.
In the first part of the talk, I will report the results of a study (joint work with Emmanuel Chemla and Benjamin Spector) where we used a priming paradigm to investigate the derivation of plural ambiguous sentences such as Amir and Manuel ate two cookies. This sentence can give rise to both distributive (i.e. two cookies each) and non-distributive (i.e. two cookies in total) interpretations. Non-distributive readings are traditionally thought to be derived by default, whereas distributive readings are thought to arise by applying some sort of distributive mechanism (Link, 1987; Champollion, 2017). Our results reveal that the distributive/non-distributive ambiguity gives rise to priming effects, suggesting that the abstract mechanism responsible for deriving plural ambiguities has a psychological correlate during comprehension.
In the second part of the talk, I will address the question of how languages partition the semantic space into lexical categories, with a specific focus on person systems, which categorize entities as a function of their role in the conversation (i.e., speaker(s), addressee(s), other(s)). Like other semantic category systems (e.g. color and kinship terms), not all ways of partitioning the person space are equally likely. I will present an artificial language learning study (joint work with Jennifer Culbertson) where we test whether the typological frequency of different person systems correlates with their learnability. We focus on the difference between 1st, 2nd and 3rd personal pronouns and test the predictions of two theories which propose different sets of features to capture the division of this space (Harley & Ritter 2002, Neeleman & Ackema 2018). Our findings provide the first experimental evidence for a feature-based theory of person systems that relies on an inherent asymmetry between the speaker and the addressee roles.