Punctuation, Prosody, and Discourse: Afterthought Vs. Right Dislocation

Abstract

In a reading production experiment we investigate the impact of punctuation and discourse structure on the prosodic differentiation of right dislocation (RD) and afterthought (AT). Both discourse structure and punctuation are likely to affect the prosodic marking of these right-peripheral constructions, as certain prosodic markings are appropriate only in certain discourse structures, and punctuation is said to correlate with prosodic phrasing. With RD and AT clearly differing in discourse function (comment-topic structuring vs. disambiguation) and punctuation (comma vs. full stop), critical items in this study were manipulated with regard to the (mis-)match of these parameters. Since RD and AT are said to prosodically differ in pitch range, phrasing, and accentuation patterns, we measured the reduction of pitch range, boundary strength and prominence level. Results show an effect of both punctuation and discourse context (mediated by syntax) on phrasing and accentuation. Interestingly, for pitch range reduction no difference between RDs and ATs could be observed. Our results corroborate a language architecture model in which punctuation, prosody, syntax, and discourse-semantics are independent but interacting domains with correspondence constraints between them. Our findings suggest there are tight correspondence constraints between (i) punctuation (full stop and comma in particular) and syntax, (ii) prosody and syntax as well as (iii) prosody and discourse-semantics.

From copula to focus, vice versa, or neither?

“…Moreover, some examples with lɛ ́ do not involve arguemerges on a residual basis, from the fact that what appears to be new, focal information is put in the beginning of the utterance and whatever is added thereafter should a fortifiori be treated as afterthought, or because there is independent evidence for such treatment. Such independent evidence could be provided by prosodic effects (Kalbertodt, Primus & Schumacher 2015). ment-predicate reversal, such as in ( 60), where the argument of the PP is the semantic predicate of a referential identity construction.…”

Typologizing nominal expressions: the noun phrase and beyond

“…This is particularly important for languages where such cases are not, or not always, distinguishable from true nominal expressions on morphosyntactic grounds, e.g., where word order is quite free and where nominals show the same morphological marking (e.g., inflecting for case, gender, number or other categories) independent of whether or not they form a discoursefunctional unit (see Reinöhl 2020b; Schultze-Berndt and Simard 2012). For example, cases involving elements in the left or right periphery may not form nominal expressions with elements inside the core clause, and are set off intonationally (Carroll 2020;Himmelmann this issue;Olsson this issue;Reinöhl 2020b;Schultze-Berndt and Simard 2012: 1025-1028; for prosodic properties in English, see Kalbertodt et al 2015). 4 In the right periphery, oft-mentioned types are afterthoughts, right dislocations, and similar phenomena.…”

The Information Structure–prosody interface in text-to-speech technologies. An empirical perspective

“…This variety ensures a minimum amount of thematicity in terms of: (i) the number of hierarchical levels of thematicity (up to three in this corpus), (ii) the presence/absence of each type of thematicity span, (iii) the position of spans within the sentence and with respect to each other, and (iv) the continuity of spans or lack of it. Given that punctuation is known to affect prosodic phrasing when reading (Kalbertodt et al 2015), a representative number of punctuation marks was also taken into account for the selection, including question marks, quotes, semicolons, and commas with different functions. In average, a sentence of our corpus contains 15 words, with a minimum of three words and a maximum of 30.…”