The constructionist approach argues that communication is central to language learning, language use, and language change. We argue that the approach provides a useful perspective on how autistic children learn language, as it anticipates variable outcomes and suggests testable predictions. First, a reduced ability and interest in tracking the attention and intentions of others should negatively impact early language development, and a wealth of evidence indicates that it does. Second, and less discussed until recently, a hyperfocus on specifics at the expense of generalizations, common among people on the spectrum, should also negatively impact language development, and recent evidence suggests this is also the case. Pace Kissine's 2021 target article, it is unsurprising that children can learn some second language from watching videos, and it is unclear how an appeal to ‘innate’ language-specific knowledge could explain the range of outcomes of individuals on the autism spectrum.