Sentence-initial Discourse Connectives, Discourse Structure and Semantics
Bonnie Webber, Rashmi Prasad · 2008
This paper focusses on differences in the properties of sentence-initial (hereafter, s-initial) coordinating conjunctions (hereafter, CCs) and s-initial discourse adverbials, and what such differences imply for discourse structure and semantics. Data are mainly from the Penn Discourse TreeBank1 (PDTB), which contains the annotation of discourse relations in 2304 articles of the Wall Street Journal corpus (Marcus et al., 1993) in terms of discourse connectives and the minimal text spans that give rise to their arguments (Prasad et al., 2008), as in Example 1: (1) Even so, according to Mr. Salmore, the ad was ”devastating ” because it raised questions about Mr. Courter’s credibility. But it’s building on a long tradition. (0041) The connective (“but”) is underlined, the first of its two arguments, ARG1, is shown in italics and the second, ARG2, is shown in boldface. The number 0041 indicates that the example comes from subsection wsj 0041 of the corpus. All annotation has been done by two annotators and adjudicated by a third (often a committee) in cases of disagreement. Only where connectives serve to connect clauses (or possibly an event-denoting nominalization, in the case of ARG1), have they been annotated. The PDTB also annotates the sense of each connective and the attribution of both connectives and their arguments, but since neither plays a role in our presentation here, we omit further discussion. As well as annotating explicit connectives, the PDTB annotates implicit connectives, which the