When people who do not speak the same language have to interact in real-time, they sometimes rely on simultaneous interpretation to understand each other. Simultaneous interpreters can listen to the speech of one person, while simultaneously translating that speech into a different language to relay it to another person. This is a difficult skill that takes years of training to master fully, an extraordinary example of speech production and comprehension under the most taxing conditions. As such, it not only presents fascinating questions about the workings of speech comprehension and production, but it also offers us a method for studying these questions.
In his dissertation, Jeroen van Paridon examined speech comprehension and production during simultaneous interpreting using a variety of methods. He proposed and tested a computational model of the temporal coordination of speech comprehension and production processes during interpreting, as well as a statistical model of the lexical and contextual factors that determine speech latencies in interpreting. In support of the work on simultaneous interpreting, he also produced novel, multilingual word embeddings for use in psycholinguistic research, as well as an exploration of the consequences of multicollinearity in commonly used language statistics.