Dr. Noah D. Goodman | Stanford University
Dr. Jenny Culbertson | University of Edinburgh
Building experimentally informed theories of typology
Human languages exhibit striking variation. At the same time, certain linguistic patterns crop up again and again, while others seem to be extremely rare. What these tantalising observations tell us about human language is one of the most contentious questions in linguistics. Do similarities between languages reflect accidents of history? A special capacity for language in humans? More general features of the human mind? Do they reflect hard-and-fast constraints on the space of possible languages? Or soft biases that influence learning and usage? Traditionally, linguists have argued for one or another of these answers based on limited sources of evidence. For example, it is common to base claims about universality on small samples of languages, case studies of how a handful of languages change over time, or examples of how individual languages are learned. In this talk, I use two case studies to highlight how behavioral experiments, targeting diverse participant populations, can be used to bring crucial empirical evidence to bear on how language is shaped (or not!) by the human linguistic and cognitive system. In the first case study, I summarise a strand of research on the role of systematicity and conceptual knowledge in shaping cross-linguistic patterns of nominal word order. In the second, I describe recent experimental work on typological trends in personal pronoun systems. In each case, I discuss how this kind of evidence can be used to improve models and theories of linguistic typology.
Dr. Jacob Andreas | Massachusetts Institute of Technology
Playing Language Games With and Like People
The last few years have seen a series of rapid developments at the intersection of machine learning and game theory. Today, systems that combine learned models of human behavior and game-theoretic decision-making techniques can compete with the best humans in games that involve bluffing, alliance-building, and long-term planning. Game-theoretic approaches have long been a staple in computational models of language change and pragmatic communication—what can this new generation of game theory techniques teach us about human language? I’ll present recent work revisiting game-theoretic models of pragmatics, offering a new game-theoretic interpretation of the Rational Speech Acts model and related probabilistic models of language comprehension. I’ll conclude by briefly discussing how these same techniques can be applied in NLP to improve the truthfulness and coherence of neural language models on question answering tasks. This is joint work with Athul Jacob, Yikang Shen and Gabriele Farina.