On universal Free Choice items

Paula Menendez-Benito 0 0 P. Menendez-Benito (&) Department of English Language and Literature, University of Gottingen , 37073 Gottingen, Germany This paper deals with the interpretation and distribution of universal Free Choice (FC) items, such as English FC any or Spanish cualquiera. Crosslinguistically, universal FC items can be characterized as follows. First, they have a restricted distribution. Second, they express freedom of choice: the sentence You can take any card conveys the information that the addressee is free to pick whichever card she chooses. Under standard assumptions, the truth conditions of sentences like You can take any card are taken to be captured by formalizations in which a universal quantifier ranging over individuals has wide scope over the possibility modal. The crucial observation that informs the account in this paper is that this type of formalization cannot capture the freedom of choice component. I will argue that in order to derive the right interaction between possibility modals and FC items, we need to add an exclusiveness condition to the standard wide scope paraphrases. The same proposal that guarantees freedom of choice will automatically account for the distribution restrictions of FC items. The formal implementation of this proposal is cast in the Hamblin semantics proposed in Kratzer and Shimoyama (Indeterminate phrases: The view from Japanese, 2002). - paraphrase.1 Both Spanish cualquiera and English any belong to this category. By uttering the Spanish sentence in (1a) or its English counterpart in (1b) I inform my addressee that all the cards are permitted possibilities for her; thus I grant her the unrestricted liberty of individual choice (Vendler 1967: 80). These sentences are often paraphrased as For every card x in this deck you can take x. Puedes coger cualquier carta de esta baraja. You can take any card from this deck. You can take any card from this deck. Universal FC items have a restricted distribution. Crosslinguistically, they are licensed in possibility sentences (see (1) above) and ruled out in episodic sentences (as in (2)) and in necessity statements like (3). These items thus seem sensitive to some property of their environment. But in view of the data in (1) through (3) it is far from obvious what this property might be: looking at the contrast between (1) and (2) one might think that the relevant property is modality (that is, that universal FC items are modality-sensitive elements). But as (3) shows, not any type of modal will do. a. *Juan cogio cualquiera de las cartas de esta baraja. Juan took-pfv. any of the cards from this deck. b. *John took any of the cards from this deck. a. *Juan tiene que coger cualquiera de las cartas de esta baraja. Juan must take any of the cards from this deck. b. *John must take any of the cards from this deck. While universal FC items seem to involve quantification, they dont behave like textbook quantifiers. As illustrated by the examples above, they have a puzzling distribution and they have obligatory wide scope with respect to possibility modals. An analysis of these items is thus likely to contribute insights into the properties of quantification in natural language. Much current research aims to provide such an analysis (see e.g., Aloni 2003, 2007a, b; Abrusan 2007a; Arregui 2006; Dayal 1998, 2004; Chierchia 2006; Farkas 2006; Giannakidou 2001; Giannakidou and Cheng 2006; Horn 2000, 2005; Jayez and Tovena 2005; Quer 1998, 2000; Sb 2001, among others). Many of these proposals share the assumption that the truth conditions of possibility sentences like (1) above can be captured by wide scope universal formalizations such as For every x such that x is a card in this deck, you may take x. A number of different ways of deriving paraphrases that essentially boil down to this structure have been proposed in the literature (see, e.g., Aloni 2003, 2007a; Dayal 1998, 2004; Chierchia 2006; Sb 2001). The crucial observation that informs the account in this paper is thatsurprisinglythis type of formalization 1 Until relatively recently the term Free Choice Item has been reserved for items of the any-type, which have a universal interpretation. Kratzer and Shimoyama (2002) have shown that some existential indefinites (e.g., German irgendein) also have a FC component. Most of the FC literature now differentiates between Existential FC items (irgendein and its kin) and Universal (any-like) FC items. cannot capture the freedom of choice component. I will argue that in order to derive the right interaction between possibility modals and FC items we need to add an exclusiveness condition to the standard wide scope universal paraphrases. Once this proposal has been implemented, both the ungrammaticality of the FC items in (2) and (3) and their obligatory wide scope in (1) will fall out automatically. The paper is organized as follows. Section 2 is devoted to capturing the FC effect. In Sect. 2.1 I argue that wide scope universal paraphrases do not guarantee freedom of choice. Section 2.2 shows how this problem can be overcome by adding an exclusiveness condition and Sect. 2.3 puts forward a compositional implementation of this idea, cast in Kratzer and Shimoyamas Hamblin semantics (Kratzer and Shimoyama 2002). Section 3 shows that the proposal developed in 2.3 accounts for the distributional restrictions of FC items exemplified above. In Sect. 4, I briefly discuss some of the questions raised by the proposal. Section 5 deals with necessity sentences that, unlike (3) above, license universal FC items (e.g., Any student must work hard). Section 6 discusses the repair strategy dubbed subtrigging in LeGrand (1975). Finally, Sect. 7 presents some issues for further research. 2 Accounting for the Free Choice effect 2.1 Wide scope universal quantification does not guarantee Free Choice One of the central questions of research on FC items of the any-type has been how to derive their universal flavor. Kadmon and Landman (1993) analyzed any as an indefinite that gets a universal-like interpretation in generic contexts. However, Dayal (1998) convincingly argued that FC any does not pattern with generic indefinites, and proposed instead that this item is a universal quantifier that ranges over possible individuals. On Sbs (2001) analysis, Scandinavian FC items also contribute a universal determiner as part of their meaning. A different view is put forward in Aloni (2003, 2007a), where the universal interpretation we see in sentences like Anyone may enter comes about because the possibility modal introduces universal quantification over propositions. Chierchia (2006) treats FC items as existential quantifiers and claims that their apparent universal force results from combining the existential assertion with a domain widening implicature (in the spirit of Kratzer and Shimoyama 2002). This list, while by no means exhaustive, illustrates the wide range of proposals that have been en (...truncated)