Self-paced graph reading and the complexity of embedded implicatures

Team: Anton Benz (SIGames, ZAS Berlin) and Maria Spychalska (U Bochum)

The aim of this collaboration is to investigate how the inferential complexity of so-called local and global scalar implicatures influences their availability. We hypothesized that their differential availability may be an effect of differences in the strategies to verify them, and the differing amount of information which they involve. The planned experiment will be implemented as a self-paced graph reading task. This means that subjects are asked to decide about the truth of sentences in situations represented by a graph about which they incrementally receive information until they feel confident to make a decision. We, first, want to find out how the complexity of the verification task influences the availability of local and global implicatures, and, second, whether the different implicatures leave a mark on processing times. One possible outcome of the study is that subjects have a strong tendency to take the first interpretation available independently of whether it is a local or a global implicature, which would indicate that the complexity of the task is a confounding parameter for experiments on the existence of embedded implicature.