Corpora in Language Acquisition Research

Corpus research forms the backbone of research on children's language development. Leading researchers in the field present a survey on the history of data collection, different types of data, and the treatment of methodological problems. Morphologically and syntactically parsed corpora allow for the concise explorations of formal phenomena, the quick retrieval of errors, and reliability checks. New probabilistic and connectionist computations investigate how children integrate the multiple sources of information available in the input, and new statistical methods compute rates of acquisition as well as error rates dependent on sample size. Sample analyses show how multi-modal corpora are used to investigate the interaction of discourse and linguistic structure, how cross-linguistic generalizations for acquisition can be formulated and tested, and how individual variation can be explored. Finally, ways in which corpus research interacts with computational linguistics and experimental research are presented.