Abstract: 

Whenever one observes a typological tendency, it is natural to ask about its source. In this talk, I explore the idea that typological tendencies arise from learning biases that become amplified over time. I will present three artificial language learning studies that investigate the existence of such biases, focusing on their potential role in explaining cross-linguistic tendencies in the domain of negation and negative dependencies.
 
I will begin by discussing a study that investigates the phenomenon known as the _nall_ lexical gap, which refers to the generalization that, across the world’s languages, the quantificational concept _not all_ (_nall_) is never lexicalized. Through a series of artificial language learning experiments, we explore whether this gap might stem from a cognitive bias against such meanings.

Next, I will turn to negative dependencies, beginning with Negative Polarity Items (NPIs)—words that frequently co-occur with negation (e.g., English _at all_). Due to their restricted distribution, learners might attribute varying interpretations to NPIs. For example, _at all_ might be understood as a universal quantifier scoping above negation or as an existential quantifier scoping below it. This study investigates how learners interpret NPIs in positive environments and whether this sheds light on the existence of such expressions in natural languages.

Finally, I will present a third study examining linguistic variability regarding how sentences with two negative elements are interpreted—whether they yield double negation or negative concord meanings. Using an artificial language learning experiment, I explore whether English speakers are sensitive to the status of negative markers when learning double negation and negative concord systems.