From roots to DPs, Brazilian Portuguese and English: a model theoretical approach to language variation

The paper develops Pires de Oliveira’s (2020, in press) model theoretical account to Brazilian Portuguese (BrP) and English, both number marking languages, ie when n is first projected, little-n, [n 0 [X]], where X is a non-categorized root, it is as a predicate, . It revises Chierchia 2010, 2014) so it might account for the difference between these two languages: BSs in English are singular predicates, whereas in BrP, they can be arguments. The proposal assumes that n 0 denotes a part-whole non-atomic lattice (Rothstein 2010, 2017), ie. without grammatical atoms. In English, n 0 attracts atomicity, generating n 1 [ AtP SG-PL [ n0 [X]]. Thus, it predicts no BSs in argument position, and coercion to mass if in such position. In BrP, [n 0 [X]] surfaces in argument position, because the nominal phrase gravitates around the “specifier”. This is the bifurcation separating these two languages. The BS in BrP conveys no grammatical information about atomicity, because there is no specifier, so no atomicity is called for. It is sub-specified for mass and count. This move allows for an unitarist approach to the nominal phrase in BrP: where number gravitates around the specifier. The conclusion explores some consequences in the domain of language variation (Lima & Rothstein 2020), and in that of semantic processing.