近期活動

[Abstract]

Adults are much less successful at learning a new language than are children. Unfortunately, debates about why this is the case have not changed much in the last half century, with roughly the same hypotheses under discussion now as in the late 1960s. Part of the difficulty is the scale of the phenomenon: people's linguistic knowledge is vast, covering tens of thousands of words and an unknown number of grammatical patterns; learning unfolds over years and in myriad contexts. Even hundreds of studies provide a woefully incomplete picture, making it difficult for researchers to agree as to what the phenomenon is, much less what explains it. I describe a series of studies that aim for (closer to) the necessary scale, including meta-analyses, massive online studies, and "Big Data" machine learning approaches. As the picture begins to clarify, some results are expected, while others suggest the need for a radically new theory. (I will not be providing a radically new theory, but I will point a few possible directions.)