Comparatives Combined with Scalar Particles: The Case of Chinese HAI

University of Pennsylvania Working Papers in Linguistics Volume 24 | Issue 1 Article 23 4-2-2018 Comparatives Combined with Scalar Particles: The Case of Chinese HAI Yi-Hsun Chen Rutgers University This paper is posted at ScholarlyCommons. https://repository.upenn.edu/pwpl/vol24/iss1/23 For more information, please contact . Comparatives Combined with Scalar Particles: The Case of Chinese HAI Abstract This paper investigates the syntax and semantics of the scalar particle HAI combined with Chinese bi comparatives. Several empirical facts are presented and discussed. First of all, two syntactic positions of HAI are identified: syntactically, the scalar particle HAI can either precede the comparative standard (HAIhigh) or follow it (HAIlow) in the Chinese bi comparative. Second, HAIlow leads to a positive inference while HAIhigh does not. Third, although many focus particles may appear in the position of HAIhigh, they are categorically banned from the position of HAIlow. Finally, HAIhigh conveys that the assertive content contravenes the speaker’s expectation. The core proposal made in this paper is the following. Syntactically, HAIhigh is an adjunct adjoined to the degree phrase while HAIlow occupies the degree head. Semantically, the scalar HAI involves two core ingredients: a scale and some presuppositional conditions based on the type of the scale. In particular, HAIhigh employs the scale of likelihood (similar to English even) and presupposes that the prejacent p is less likely than its alternative ¬p: the negation of the prejacent. By contrast, HAIlow takes the scale provided by gradable predicates and presupposes that both the comparative target and the comparative standard are ordered above the contextual standard of the scale. Finally, it is proposed that the semantics of the scalar HAI is constrained by its syntactic position: while HAIhigh operates on the domain of propositions based on the scale of likelihood, HAIlow on the domain of degrees based on the dimension of gradable predicates. This working paper is available in University of Pennsylvania Working Papers in Linguistics: https://repository.upenn.edu/pwpl/ vol24/iss1/23 Comparatives Combined with Scalar Particles: The Case of Chinese HAI Yi-Hsun Chen* 1 Introduction In Chinese, the particle HAI has three different uses: (i) an aspectual use, similar to English still; (ii) an additive use, similar to English also; (iii) a scalar use, similar to English even (e.g., Liu 2000, Liu et al. 2001, Yang 2017).1 The first two uses are exemplified in (1). Informally, on the aspectual reading, (1) presupposes that there was some time in the past abutting the utterance time such that Zilu liked Xiaomei during that time.2 On the additive reading, suppose the focus associate is Xiaomei, (1) presupposes that there is someone in the previous discourse such that Zilu like her, in addition to Xiaomei. (1) Aspectual and Additive Use a. Zilu HAI xihuan Xiaomei. Zilu still/ also like Xiaomei ‘Zilu still/ also likes Xiaomei.’ This paper focuses on the scalar use of HAI in the Chinese bi comparative, as illustrated in (2). For the purposes of discussion, I refer to the scalar HAI in the case of (2a) as HAIhigh and in the case of (2b) as HAIlow. Intuitively, there are two differences between (2a) and (2b). First, the latter, but not the former, leads to a positive inference: both Zilu and Lisi are presupposed to be above the contextual standard of tallness in (2b).3,4 Second, HAIhigh is apparently outside the degree phrase while HAIlow is part of the degree phrase, as the latter syntactically intervenes between the gradable predicate and the comparative standard.5 (2) Scalar Use a. Zilu HAI bi Lisi Zilu HAI than Lisi ‘Zilu is taller even than Lisi.’ gao. tall * I am grateful to Mark Baker, Simon Charlow, Huiyu Huang, Peiyi Hsiao, Jess Law, Ang Li, Haoze Li, Lydia Newkirk, Chris Oakden, Shuhao Shih, Livia Carmargo Souza, Chingyu Yang for constructive suggestions and comments. I am also grateful to three reviewers of PLC41 and the audiences for helpful comments and discussions. Of course, all errors are mine. 1 Chinese HAI is not the only particle showing the three-way ambiguity. German noch is another particle well-known for having the three uses: the aspectual use, the additive use and the scalar use. See Krifka (1999, 2000), Umbach (2009, 2012) and recently Beck (2016) for discussion of the meanings of German noch. 2 See Krifka (2000), Ippolito (2007) and Greenberg (2009) for discussion of the meaning of English still. 3 The rise of the positive inference in a comparative construction is surprising because a comparative by itself does not require both the comparative target and comparative standard to be above the contextual standard of the dimension provided by the gradable predicate. Consider (i) and (ii). In English, (i) does not require both Adam and Bill to be above the contextual standard of tallness. By contrast, even (similar to Chinese HAIlow) triggers a positive inference in (ii): both Adam and Bill are above the contextual standard of tallness. (i) Adam is taller than Bill. (ii) Adam is even taller than Bill. 4 According to Umbach (2009), German noch does not necessarily trigger a positive inference when it combines with a comparative (in the terminology of Umbach: the comparatives use of noch). In this respect, German noch differs from Engish even and Chinese HAIlow. Furthermore, according to Umbach (2009), the comparative use of German noch is not related to the scale of likelihood in its semantics. In this respect, German noch differs from Chinese HAIhigh and again from the canonical view of English even. However, see Greenberg (2016) for arguments against the traditional likelihood analysis of English even. 5 A terminological note here: I use comparative standard to refer to the standard of comparison in the comparative and comparative target to refer to the individuals/ objects that are compared with the comparative standard. For instance, Zilu is the comparative target and Lisi is the comparative standard in (2). U. Penn Working Papers in Linguistics, Volume 24.1, 2018 200 YI-HSUN CHEN b. Zilu bi Lisi HAI Zilu than Lisi HAI ‘Zilu is even taller than Lisi.’ gao.6 tall In Section 2, I provide empirical data corroborating the two differences. To anticipate, the core of my proposal is outlined below. (3) The syntax and semantics of HAIhigh a. [DegP HAIhigh [DegP Subject [Deg’ bi-phrase [Deg’[Deg [AP [A ]]]]]]] b. ∥HAIhigh∥  = λC<st,t>..λp<s,t>. λw<s>.: $q[qÎ C Ù q = ¬p Ù p <likely q]. p(w) The prejacent p is less likely to be true than its alternative q in w, with respect to a given context c; When defined, the prejacent p is true in w. (4) The syntax and semantics of HAIlow a. [DegP Subject [Deg’ bi-phrase [Deg’[Deg HAIlow [AP [A ]]]]]] b. ∥HAIlow∥  = λC<d,t>..λG<e, d>.λy<e>.λx<e>.: $d[d Î C Ù d ≥ ds Ù G(x) > d ˄ G(y) > d]. G (x) > G (y) The degrees to which individual x and individu (...truncated)