The last twenty years have witnessed the development of promising methodologies and new paradigms that have brought substantial findings and have changed our views on early language acquisition. Focusing on early comprehension, this article is mainly devoted to a review of these new paradigms analyzing their benefits and limits. One of the main challenges is the development of reliable on-line behavioural methods coupled with neurophysiological data in cross-sectional and longitudinal studies. Another challenge is the extension of these new paradigms within a powerful crosslinguistic perspective. The second part of the article focuses on some advances in different domains directly linked to new methodologies: the evaluation of task dependence in early syntactic comprehension, some new insights on production/comprehension asymmetries and the predictive value of speed of processing language in two-year-olds for language and cognitive abilities in later childhood.