In this commentary on the Special Issue, I will address the question of what memory spans measure concerning language, as language has, at least, a linear and a hierarchical dimension. I suggest that if anything what is measured has to do with the linear dimension. Then, I will discuss the welcome results on bilingual children with language problems and reflect on what it means that working memory measures seem to be more relevant for bilingual children than for monolingual ones. I will tentatively suggest that what is at stake is not memory per se, but inhibitory skills.