Quantification and focus in Negative Concord Balázs Surányi Eötvös University, Budapest suranyi@nytud.hu 1. Introduction Negative Concord (NC) appears to be a cross-linguistically varied phenomenon, despite initial optimistic hopes. 1 Studies have revealed that n-words may be assigned diverging interpretations in different languages. In some languages they carry logical negation, while in others they do not. Apparently, in some languages they are interpreted as universals (e.g. Greek, cf. Giannakidou 1998), in others as existentially quantified, and proposals have been offered in which their quantificational force varies as a function of syntactic context. This paper examines Hungarian n-words and advocates the view that Hungarian is a truly hybrid language type with respect to a typology of NC in both of these regards: (i) n- words can be negative or non-negative as an option, and (ii) they can be interpreted as universally quantified and also as existentially quantified. Importantly, the two choices are independent of each other. The simultaneous presence of these options within the same language (confirming a prediction of Giannakidou 2000) furnishes solid evidence that the same options are available within a cross-linguistic typology of Negative Concord. The paper is structured into two main parts. First, I refute two—opposing—views of the negativity of Hungarian n-words, each of which classifies Hungarian as belonging to what can be referred to as a pure type of Negative Concord. According to Puskás (1998, 2000) n- words in this language invariably carry logical negation, while according to É.Kiss (1998, 2002) and Olsvay (2000), they are non-negative. I argue that these accounts mischaracterise the role played by a negative particle optionally appended to n-words, and that the correct approach factors Hungarian n-words into two morpho-syntactic classes: one that is semantically negative and another that is non-negative. Second, I explore the issue of the quantificationality of n-words. I demonstrate that both the negative and the non-negative n-word varieties can be interpreted either as universally or as existentially quantified. In exploring the distribution of the two interpretational options, I argue that the ambiguity is one between the presence and lack of a universal quantifier, the latter situation resulting in a non-quantified Heimian indefinite interpretation, subject to existential closure. A well-definable gap in the coverage of the mapping from the syntactic positions occupied by the n-words to their quantificational status is eliminated by demonstrating that n-words can (optionally) be fronted by syntactic focalization due to a morphosyntactically reflected ‘even’ component in their semantics. This result provides strong confirmation for treatments of NPI put forward by Lee (1993), Lahiri (1998) and Horn (2000). I conclude with spelling out the main consequence of these findings for a general typology of Negative Concord, and their connection to the Jespersen Cycle. 2. The negativity of Hungarian n-words 2.1 The patterns I first review the basic syntactic patterns in Hungarian Negative Concord clauses. N-words in this language may remain postverbal if there is an overt negation element in the clause, cf. (1a). 2 N-words come in two flavours: they may or may not be modified by the particle sem. I will be using the term s-words to refer specifically to unmodified, bare n-words (due to the