An increscent literature on bilingual development expamines relationships between language exposure and learning outcomes. Vocabulary size and pace of grammar learning have been asserted to be causally linked to amounts or types of exposure to each language. Strong claims are made about the role of exposure on bilingual outcomes. Some researchers posit a unique learning result: a ‘weak language’. This critical review utters reasons for scepticism that quantity or quality of exposure alone will clarify findings. Central constructs are not well defined; unfitting research methods have been employed; the right kind of data is not debated. Crucially, authors lie on the notion of language itself, switching between cognitive and environmental viewpoints. Both are required to interpret bilingual behaviours yet play distinct roles in the construction of learner grammars.