Emphatic Stress as Epistemic Conflict: A case study of Mandarin Chinese

This paper examines the semantics of sentence-final stress in Mandarin Chinese and gives a uniform account of the distribution and interpretation of emphatic stress. A phonetic production study shows that the use of sentence-final stress by Mandarin speakers is sensitive to the evidentiality of the context. A semantic rating study shows that listeners judge sentences with evidenced contexts as more natural when they are accompanied by sentence-final stress. Based on these results, we argue that a sentence-final stress gives rise to an alternative proposition which is obtained by switching the polarity of the sentence, and that the sentence-final stress indicates an epistemic conflict between the speaker’s belief and the incoming proposition at issue. 1 Prosodic Stress in Mandarin Chinese In Mandarin Chinese, prosodic stress on a constituent can indicate an emphatic focus (c.f., Liu and Xu [2005], Peng et al. [2006]). In (1), the subject NP Weili receives emphatic stress (marked as es), which gives rise to a narrow focus on Weili ; that is, (1) indicates that among alternative individuals, it is Weili that sells bacon, not other people. Phonetically, the emphatic stress is realized via local expansion of the pitch range of focused syllables as well as post-focus compression, as depicted in Figure 1 (taken from Peng et al. [2006]). Semantically, emphatic stress gives rise to alternative propositions. In the case of constituent stress like (1), the alternative propositions are obtained by replacing the focused constituent with alternative constituents of the same semantic type (e.g., type e in (1), see Rooth [1985]). For instance, we obtain the following alternative set for (1): Alt(![F Weili] sells bacon") = {!Weili sells bacon", !Zhansang sells bacon", !Xiaomei sells bacon", ...}. Further pragmatic strengthening will yield exhaustification [see Schulz and van Rooij, 2006]. (1) [esWei51li51] Weili mai51 sell la51rou51. bacon ‘WEILI sells bacon’ (It is Weili, not others, that sells bacon.) (adapted from Peng et al. [2006]) (a) Narrow focus on WeiLi (b) “Out of the blue” statement Fig. 1. Pitch expansion due to constituent stress in Mandarin Chinese [Peng et al, 2006; p. 52] The same pitch expansion can occur sentence-finally. In (2), stress (marked as sfs) falls on the last syllable of a falling positive declarative, and in (3), a rising negative declarative. This paper investigates the phonetics and semantics of these types of sentences with sentence-final stress. Semantically, introspectionbased data suggests that when the verb-final syllable bu51 is pronounced with stress, the utterance indicates that the speaker had a previous belief toward a proposition with the opposite polarity. More specifically, final stress in a falling declarative like (2) gives an impression of the speaker’s surprise, while final stress in a rising declarative like (3) indicates the speaker’s incredulity about the embedded proposition. (2) Ni21 You ping35shi35 usually pao21[sfsbu51]! jogging ‘You usually jog!’ (Falling Positive Declarative) (I believed that you don’t usually jog.) (3) Ni21 You ping35shi35 usually bu51 NEG pao21[sfsbu51]? jogging ‘You don’t usually jog?’ (Rising Negative Declarative) (I believed that you usually jog.) Building on the discussion of the focus effect of constituent stress observed by Peng et al. [2006], we propose that the emphatic stress is licensed only if the alternatives are salient in the context. In the case of sentence-final stress, stress indicates a focus-marking on the polarity of the sentence; hence the alternative proposition is obtained by switching the polarity of the uttered sentence. The alternative proposition is salient because it is part of the speaker’s previous belief. In sum, constituent stress and sentence-final stress can both be characterized as triggering salient alternatives. When does the speaker have a belief toward a particular proposition? The most plausible situation is when the speaker has some reason to believe the proposition, i.e., evidence which supports the proposition. That is, the strength of evidence determines the speaker’s epistemic states and therefore influences the change of belief. Therefore, our proposal predicts that a sentence uttered with sentence-final stress is more natural when the speaker has evidence for the alternative proposition and thus believes it. Our phonetic production study shows that evidential intonation indeed exhibits characteristics of focal stress in Mandarin Chinese. The semantic naturalness rating study also shows that Mandarin listeners judge the sentences with evidence more natural when they are accompanied by sentence-final stress. We thus conclude that sentence-final stress indicates an epistemic conflict between the proposition at issue and the speaker’s previous belief. 2 A production study We started with a production study to investigate whether Mandarin Chinese speakers use sentence-final stress to reflect the evidentiality of context. Liu and Xu [2005] show that focal stress in Mandarin Chinese is phonetically characterized by an elevated and expanded pitch range. The current production study aimed to investigate whether evidentiality would induce a similar type of focal stress.