Carroll's paper (Carroll) is most welcomed and shows that the weaker language is still to be explored from various points of views. Over twenty years after the publication of Schlyter's paper, “The weaker language in bilingual Swedish-French children”, not only is the topic of the weaker language of simultaneous bilinguals still a clearly under-researched area, but it is also, as Carroll shows, a crucial domain of research in the discussion of properties and constraints of the human Language Making Capacity (LMC).

However, although the interest in the weaker language has been quite weak ever since, it has not been as methodologically and theoretically weak as it appears in Carroll's article; in fact, just by considering a few more of some of the recent studies, we may shed light on the role of exposure and input in the development and ultimate attainment in the weaker language.

Studies by Kupisch and colleagues speak well for themselves, as for what concerns the definition of central constructs, research methods and the right kind of data – which according to Carroll lack in the literature on the role of input and exposure. Kupisch (Reference Kupisch2012, Reference Kupisch2014), Kupisch, Akpinar and Stöhr (Reference Kupisch, Akpinar and Stöhr2013), and Stöhr, Akpinar, Bianchi and Kupisch (Reference Stöhr, Akpinar, Bianchi, Kupisch, Braunmüller and Gabriel2012) have shown that adult simultaneous bilinguals behave like monolinguals in their majority language – i.e., the language of the place were they were brought up – but differently in their minority language – the language of one of the parents and other relatives – and, also, that majority language bilingual speakers outperform minority language bilingual speakers in various grammatical domains; so much so that the results in the weaker language are qualitatively comparable to L2 results.

To give an example, Kupisch (Reference Kupisch2012), testing readings, of definite DPs and the ability to correct indefinite DPs and bare NPs in generic contexts by adult German–Italian simultaneous bilinguals and German L2 learners of Italian, found clear differences between Italian as the weaker and as the stronger language and similarities between Italian as L2 and the weaker language. This result was also observed in Kupisch (Reference Kupisch2014) on the placement of adjectives in adult German–Italian simultaneous bilinguals and L2 learners of Italian, which shows clear differences between Italian as the weaker and as the stronger language, and similarities between Italian as L2 and the weaker language. Apparently, input and exposure are important for at least these domains.

However, Granfeldt and colleagues have shown that input and exposure are important for some grammatical domains but not for others. In Ågren, Grandfeldt and Thomas (Reference Ågren, Granfeldt and Thomas2014), on the role of input and exposure in simultaneous and successive bilingual acquisition of various grammatical domains of French, the children who were exposed to French earlier than at the age of 3,5 years followed the same acquisition path as monolinguals; while those with an age of onset above 6 followed the adult L2 trajectory in the acquisition of finiteness. In Granfeldt (under review) on the acquisition of different aspects of grammatical gender, there is a strong correlation between input conditions and gender assignment, distinguishing child L2 learners from 2L1 learners qualitatively by the errors they make, while for gender concord, a combination of factors are responsible for the acquisition path.

To conclude, the studies I have mentioned show that there is no clear cut off point of all phenomena at a certain age, but that for some structures input and exposure are more important than for others. The result and conclusion of Schlyter (Reference Schlyter, Hyltenstam and Viberg1993) turns out to be confirmed at least for several areas of grammar, for child and adult weaker language speakers; so, at the end of the day, at least in my reading of the available literature – which differs from Carroll's – there's nothing that contradicts Schlyter (Reference Schlyter, Hyltenstam and Viberg1993)’s initial hypothesis: that the weaker language of simultaneous bilinguals and adult L2 learners follow the same developmental trajectories.