Prosodic and structural variability in free word order language discourse

Prosodic variability;word order

Luchkina, Tatiana Vyacheslavovna

关键词: Prosodic variability;    word order;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/93033/LUCHKINA-DISSERTATION-2016.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Independently of the modality of presentation (written or auditory), human processing of discourse obligatorily involves monitoring relative information prominence which reflects how important information is in discourse, and thereby determines the perceptual impact it makes on the speaker and the listener. In spoken language use, relative information prominence is expressed it by means of morphology, structural organization of information across an utterance, and by prosodic means (Morgan, Meier, & Newport 1987, Stolterfoht, Friederici, Alter, & Steube 2007, Watson 2010). To illustrate, Hindi speakers may use special morphological prominence markers, bound morphemes 'hii' and 'bhii', which attach to words that the listener is likely to identify as prominent (Luchkina, Puri, Jyothi, Cole 2015). In Hungarian and Hindi, speakers place the prominent word in the pre-verbal position in a sentence or phrase, which presents a designated location for prominent (focused) information in these languages (Genzel and Kügler 2010, Féry 2013). In English, it is the utterance-final or, else, the most prosodically prominent word in a sentence or phrase that is likely to be identified as prominent (Ladd, 2008, Watson 2010).This thesis examines the use of acoustic-prosodic cues and constituent ordering in the expression of relative information prominence and the way it affects perception (as perceived prominence) in Russian, a free word order language, by empirically testing the "dual route" model of expressing prominence in discourse. This model presupposes (1) structural "packaging" of information, evident from the linear ordering of words in an utterance such that words communicating relatively more accessible and therefore less salient information precede words communicating less accessible and therefore more salient information, and (2) varying magnitude of acoustic-prosodic parameters in a controlled way such that prominent information bears greater perceptual salience in speech. Speech production and comprehension experiments described below test whether these routes, structural and acoustic-prosodic, are used independently or together in the encoding of information prominence. Russian is chosen as the test case because it allows but does not require surface reordering of sentential constituents for information structural purposes and exhibits distinctions in prosodic prominence among the constituents of a sentence (Sekerina 2003, Slioussar 2011a, b, Svetozarova 1998).To examine how prosodic and structural cues are utilized during the off-line and the online processing of discourse in Russian, the following research objectives are pursued. In Study 1 (see Chapter 2 of the present version), the distribution of structural and acoustic-prosodic variability in read discourse is examined in association with two well-known prominence scales: distinctions in the information status of a discourse referent and animacy of a discourse referent (in conjunction with grammatical function of the corresponding lexical word). In Study 2 (see Chapter 3 of the present thesis), relative contribution and perceptual validity of linearization prominence cues and acoustic-prosodic prominence cues is examined using perceived prominence ratings solicited from linguistically-naïve native speakers of Russian. In Study 3 (see Chapter 4 of the present thesis), processing costs associated with these prominence cues are gauged using probe recognition response times obtained during online comprehension of discourse fragments with experimentally controlled variation in word order and acoustic-prosodic expression.