Arndt Riester

Referential and lexical givenness: Semantic, prosodic and cognitive aspects

The main objective of the paper is to show that for an adequate analysis of an item’s information status in spoken language two levels of givenness have to be investigated: a referential and a lexical level. This separation is a crucial step towards our goal to arrive at the best possible classification of nominal expressions occurring in natural discourse which reflects our understanding of all important aspects related to an item’s givenness or novelty (“information status”). For that purpose, we first introduce our motivation for this division which stems from the observation that both levels of information status have an influence on an item’s prosodic marking. Section 2 presents basic concepts including the cognitive dimensions which have been discussed in the literature on givenness, in particular the role of knowledge, consciousness and (un)importance. In Section 3, we give an overview of influential proposals to classify and order the various concepts. In doing so, we explain and relate the most important concepts that play a role in classifying nominal expressions: coreference, bridging, inference, hearer knowledge, indexicality, embeddedness. A number of paradoxes and inconsistencies of the presented approaches are discussed, which can be resolved by the insight that there are two different notions of givenness which apply to expressions of different syntactic-semantic type. We refer to them as referential givenness (rgivenness) and lexical givenness (l-givenness), and provide the semantic basis for these two concepts. In Section 4, we propose a fine-grained two-level annotation scheme for the analysis of an item’s information status, called the RefLex scheme. Section 5 takes a closer look at the prosody of the proposed labels in combination and develops a couple of hypotheses which are tested in two small exemplary corpora of spontaneous and read German speech.

A Discourse Information Radio News Database for Linguistic Analysis

In this paper we present DIRNDL, an annotated corpus resource comprising syntactic annotations as well as information status labels and prosodic information. We introduce each annotation layer and then focus on the linking of the data in a standoff approach. The corpus is based on data from radio news broadcasts, i.e. two sets of primary data: spoken radio news files and a written text version which sometimes deviates from the actual spoken data. We utilize a generic relational database management system to bridge the gap between the deviating primary data as well as between the different properties of the annotation levels. We show how the resource can support data extraction concerning the interface between information status, syntax and prosody.

Incorporating Information Status into Generation Ranking

We investigate the influence of information status (IS) on constituent order in German, and integrate our findings into a loglinear surface realisation ranking model. We show that the distribution of pairs of IS categories is strongly asymmetric. Moreover, each category is correlated with morphosyntactic features, which can be automatically detected. We build a loglinear model that incorporates these asymmetries for ranking German string realisations from input LFG F-structures. We show that it achieves a statistically significantly higher BLEU score than the baseline system without these features.

Frequency Matters: Pitch Accents and Information Status

This paper presents the results of a series of experiments which examine the impact of two information status categories (given and new) and frequency of occurrence on pitch accent realisations. More specifically the experiments explore within-type similarity of pitch accent productions and the effect information status and frequency of occurrence have on these productions. The results indicate a significant influence of both pitch accent type and information status category on the degree of within-type variability, in line with exemplartheoretic expectations.

Squiggly issues: alternative sets, complex DPs, and intensionality

In this paper, we investigate a number of long-standing issues in connection with (i) focus interpretation and its interrelation with complex definite descriptions, and (ii) the intensional properties of sentences with focus constituents. We revitalize the use of Rooths (1992) ∼ operator, clarify its definition as an anaphoric operator, discuss the principles that govern its placement in logical forms and show how it can be succesfully employed to replace the notion of Krifkas (2006) focus phrases. Finally, we argue that a proper view of the intensional dimension of retrieving the antecedent sets required by the operator can account for problems relating to the intensionality of sentences with focus sensitive operators that are discussed by Beaver & Clark (2008).

Using prosodic annotations to improve coreference resolution of spoken text

This paper is the first to examine the effect of prosodic features on coreference resolution in spoken discourse. We test features from different prosodic levels and investigate which strategies can be applied. Our results on the basis of manual prosodic labelling show that the presence of an accent is a helpful feature in a machine-learning setting. Including prosodic boundaries and determining whether the accent is the nuclear accent further increases results.