This paper is mainly concerned with two points: The first one is a better theoretical foundation of the interpretation of acceptability judgment tasks (AJTs) in studies on Brazilian Portuguese (BrP) bare nominals (BNs). I draw on Bader & Häussler's (2010) model which is based on Signal Detection Theory and show that an (explicitly or implicitly) binary approach to AJTs on BrP BNs fails to capture the whole picture. This is exemplified by contrasting the two AJT studies on specific and definite BNs presented in this paper with other experimental approaches to BrP BNs. The second concern is the status of these rather marginal forms in BrP. It will be claimed that only an approach combining different empirical methods can give a sufficiently clear picture. In order to support this claim, a third experiment, namely an elicitation task will be presented and discussed.
IntroductionAt least since their appearance as counter-evidence to Chierchia's (1998) wellknown Nominal Mapping Parameter in Schmitt & Munn (1999) 2 , Brazilian Portu-1 This work was financed by the Deutsche Forschungsgemeinschaft, SFB 833, project C3. I would like to thank Sam Featherston, Oliver Bott, Fabian Schlotterbeck and Janina Rado for their support regarding technical implementation, statistical analysis and much constructive feedback. Two anonymous reviewers helped to improve the paper at several critical points. I am also grateful to all the people in Brazil who made the data collection possible, special thanks to José Simões da Silva, Célia Regina dos Santos Lopes and Bruno Festas. The interpretation of the results as well as remaining errors are completely my responsibility.2 For a recent discussion of Schmitt & Munn's claim with regard to "semantic parameterization" and the role of the BrP data as counter-evidence to Chierchia (1998), see Dayal (2011 and references therein. 2 guese (BrP) BNs have received a considerable degree of attention from researchers interested in the syntax and semantics of the nominal domain. This is due to the fact that BrP, which has a full-fledged Romance-style article paradigm, at the same time allows for plural as well as singular BNs in argument position in surprisingly many contexts. In the words of Braga et al. (2010: 75f): "[…] the determiner phrase […] ranges from overtly definite and indefinite phrases, both singular and plural, to a complete system of bare noun phrases in argument position: bare singulars […], bare plurals […] and bare mass […]". The investigation of this topic has led to a very dynamic and productive but also controversial discussion during the last decade, especially in the case of the so-called "bare singulars" (BSs) 3 , to which this paper is mainly devoted, as well. Examples (1) and (2) contrast a BS with a bare mass noun:(1) Professor trabalha muito.Teacher works much 'Teachers work a lot.' (Müller & Oliveira 2004: 12) (2) Leite faz bem pra saúde. milk makes well for health 'Milk is healthy. ' (Braga et al. 2010: 76) While (2) is not surprising at all from a Roman...