Although diagrams are fundamental to Rhetorical Structure Theory, their interpretation has received little in-depth exploration. This paper presents an algorithmic approach to accessing the meaning of these diagrams. Three algorithms are presented. The first of these, called reenactment, recreates the abstract process whereby structures are created, following the dynamic of coherence development, starting from simple relational propositions, and combing these to form complex expressions which are in turn integrated to define the comprehensive discourse organization. The second algorithm, called composition, implements Marcu’s strong nuclearity assumption. It uses a simple inference mechanism to demonstrate the reducibility of complex structures to simple relational propositions. The third algorithm, called compress, picks up where Marcu’s assumption leaves off, providing a generalized fully scalable procedure for progressive reduction of relational propositions to their simplest accessible forms. These inferred reductions may then be recycled to produce RST diagrams of abridged texts. The algorithms described here are useful in positioning computational descriptions of rhetorical structures as discursive processes, allowing researchers to go beyond static diagrams and look into their formative and interpretative significance.

Good scientific writing makes use of specific sentence and paragraph structures, providing a rich platform for discourse analysis and developing tools to enhance text readability. In this vein, we introduce SciPara, a novel dataset consisting of 981 scientific paragraphs annotated by experts in terms of sentence discourse types and topic information. On this dataset, we explored two tasks: 1) discourse category classification, which is to predict the discourse category of a sentence by using its paragraph and surrounding paragraphs as context, and 2) discourse sentence generation, which is to generate a sentence of a certain discourse category by using various contexts as input. We found that Pre-trained Language Models (PLMs) can accurately identify Topic Sentences in SciPara, but have difficulty distinguishing Concluding, Transition, and Supporting Sentences. The quality of the sentences generated by all investigated PLMs improved with amount of context, regardless of discourse category. However, not all contexts were equally influential. Contrary to common assumptions about well-crafted scientific paragraphs, our analysis revealed that paradoxically, paragraphs with complete discourse structures were less readable.

This paper presents evidence for an effect of genre on the use of discourse connectives in argumentation. Drawing from discourse processing research on reasoning based structures, we use fill-mask computation to measure genre-induced expectations of argument realisation, and beta regression to model the probabilities of these realisations against a set of predictors. Contrasting fill-mask probabilities for the presence or absence of a discourse connective in baseline and finetuned language models reveals that genre introduces biases for the realisation of argument structure. These outcomes suggest that cross-domain discourse processing, but also argument mining, should take into account generalisations about specific features, such as connectives, and their probability related to the genre context.

We present a pipeline for multi-lingual Shallow Discourse Parsing. The pipeline exploits Machine Translation and Word Alignment, by translating any incoming non-English input text into English, applying an English discourse parser, and projecting the found relations onto the original input text through word alignments. While the purpose of the pipeline is to provide rudimentary discourse relation annotations for low-resource languages, in order to get an idea of performance, we evaluate it on the sub-task of discourse connective identification for several languages for which gold data are available. We experiment with different setups of our modular pipeline architecture and analyze intermediate results. Our code is made available on GitHub.

Discourse segmentation received increased attention in the past years, however the majority of studies have focused on written genres and with high-resource languages. This paper investigates discourse segmentation of a Taiwan Southern Min spontaneous speech corpus. We compare the fine-tuning a Language Model (LLM using two approaches: supervised, thanks to a high-quality annotated dataset, and weakly-supervised, requiring only a small amount of manual labeling. The corpus used here is transcribed with both Chinese characters and romanized transcription. This allows us to compare the impact of the written form on the discourse segmentation task. Additionally, the dataset includes manual prosodic breaks labeling, allowing an exploration of the role prosody can play in contemporary discourse segmentation systems grounded in LLMs. In our study, the supervised approach outperforms weak-supervision ; character-based version demonstrated better scores compared to the romanized version; and prosodic information proved to be an interesting source to increase discourse segmentation performance.

The identification of political actors who put forward claims in public debate is a crucial step in the construction of discourse networks, which are helpful to analyze societal debates. Actor identification is, however, rather challenging: Often, the locally mentioned speaker of a claim is only a pronoun (“He proposed that [claim]”), so recovering the canonical actor name requires discourse understanding. We compare a traditional pipeline of dedicated NLP components (similar to those applied to the related task of coreference) with a LLM, which appears a good match for this generation task. Evaluating on a corpus of German actors in newspaper reports, we find surprisingly that the LLM performs worse. Further analysis reveals that the LLM is very good at identifying the right reference, but struggles to generate the correct canonical form. This points to an underlying issue in LLMs with controlling generated output. Indeed, a hybrid model combining the LLM with a classifier to normalize its output substantially outperforms both initial models.

Writing research articles is crucial in any academic’s development and is thus an important component of the academic discourse. The Introduction section is often seen as a difficult task within the research article genre. This study presents two metrics of rhetorical moves in academic writing: step-n-grams and lengths of steps. While scholars agree that expert writers follow the general pattern described in the CARS model (Swales, 1990), this study complements previous studies with empirical quantitative data that highlight how writers progress from one rhetorical function to another in practice, based on 50 recent papers by expert writers. The discussion shows the significance of the results in relation to writing instructors and data-driven learning.

With the raise of large language models (LLMs), different evaluation methods, including probing methods, are gaining more attention. Probing methods are meant to evaluate LLMs on their linguistic abilities. However, most of the studies are focused on morphology and syntax, leaving discourse research out of the scope. At the same time, understanding discourse and pragmatics is crucial to building up the conversational abilities of models. In this paper, we address the problem of probing several models of discourse knowledge in 10 languages. We present an algorithm to automatically adapt existing discourse tasks to other languages based on the Universal Dependencies (UD) annotation. We find that models perform similarly on high- and low-resourced languages. However, the overall low performance of the models’ quality shows that they do not acquire discourse well enough.

Discourse relation classification within a multilingual, cross-framework setting is a challenging task, and the best-performing systems so far have relied on monolingual and mono-framework approaches.In this paper, we introduce transformer-based multilingual models, trained jointly over all datasets—thus covering different languages and discourse frameworks. We demonstrate their ability to outperform single-corpus models and to overcome (to some extent) the disparity among corpora, by relying on linguistic features and generic information about the nature of the datasets. We also compare the performance of different multilingual pretrained models, as well as the encoding of the relation direction, a key component for the task. Our results on the 16 datasets of the DISRPT 2021 benchmark show improvements in accuracy in (almost) all datasets compared to the monolingual models, with at best 65.91% in average accuracy, thus corresponding to a 4% improvement over the state-of-the-art.

Question Generation (QG), the process of generating meaningful questions from a given context, has proven to be useful for several tasks such as question answering or FAQ generation. While most existing QG techniques generate simple, fact-based questions, this research aims to generate questions that can have complex answers (e.g. “why” questions). We propose a data augmentation method that uses discourse relations to create such questions, and experiment on existing English data. Our approach generates questions based solely on the context without answer supervision, in order to enhance question diversity and complexity. We use an encoder-decoder trained on the augmented dataset to generate either one question or multiple questions at a time, and show that the latter improves over the baseline model when doing a human quality evaluation, without degrading performance according to standard automated metrics.

This paper proposes a classification model for single label implicit discourse relation recognition trained on soft-label distributions. It follows the PDTB 3.0 framework and it was trained and tested on the DiscoGeM corpus, where it achieves an F1-score of 51.38 on third-level sense classification of implicit discourse relations. We argue that training on soft-label distributions allows the model to better discern between more ambiguous discourse relations.

The ARRAU corpus is an anaphorically annotated corpus designed to cover a wide variety of aspects of anaphoric reference in a variety of genres, including both written text and spoken language. The objective of this annotation project is to push forward the state of the art in anaphoric annotation, by overcoming the limitations of current annotation practice and the scope of current models of anaphoric interpretation, which in turn may reveal other issues. The resulting corpus is still therefore very much a work in progress almost twenty years after the project started. In this paper, we discuss the issues identified with the coding scheme used for the previous release, ARRAU 2, and through the use of this corpus for three shared tasks; the proposed solutions to these issues; and the resulting corpus, ARRAU 3.

This study introduces an approach for evaluating the importance of signals proposed by Das and Taboada in discourse parsing. Previous studies using other signals indicate that discourse markers (DMs) are not consistently reliable cues and can act as distractors, complicating relations recognition. The study explores the effectiveness of alternative signal types, such as syntactic and genre-related signals, revealing their efficacy even when not predominant for specific relations. An experiment incorporating RST signals as features for a parser error / success prediction model demonstrates their relevance and provides insights into signal combinations that prevents (or facilitates) accurate relation recognition. The observations also identify challenges and potential confusion posed by specific signals. This study resulted in producing publicly available code and data, contributing to an accessible resources for research on RST signals in discourse parsing.

Recent language models have significantly boosted conversational AI by enabling fast and cost-effective response generation in dialogue systems. However, dialogue systems based on neural generative approaches often lack truthfulness, reliability, and the ability to analyze the dialogue flow needed for smooth and consistent conversations with users. To address these issues, we introduce GroundHog, a modified BART architecture, to capture long multi-grained inputs gathered from various factual and linguistic sources, such as Abstract Meaning Representation, discourse relations, sentiment, and grounding information. For experiments, we present an automatically collected dataset from Reddit that includes multi-party conversations devoted to movies and TV series. The evaluation encompasses both automatic evaluation metrics and human evaluation. The obtained results demonstrate that using several linguistic inputs has the potential to enhance dialogue consistency, meaningfulness, and overall generation quality, even for automatically annotated data. We also provide an analysis that highlights the importance of individual linguistic features in interpreting the observed enhancements.

Discourse analysis plays a crucial role in Natural Language Processing, with discourse relation prediction arguably being the most difficult task in discourse parsing. Previous studies have generally focused on explicit or implicit discourse relation classification in monologues, leaving dialogue an under-explored domain. Facing the data scarcity issue, we propose to leverage self-training strategies based on a Transformer backbone. Moreover, we design the first semi-supervised pipeline that sequentially predicts discourse structures and relations. Using 50 examples, our relation prediction module achieves 58.4 in accuracy on the STAC corpus, close to supervised state-of-the-art. Full parsing results show notable improvements compared to the supervised models both in-domain (gaming) and cross-domain (technical chat), with better stability.

Topics play an important role in the global organisation of a conversation as what is currently discussed constrains the possible contributions of the participant. Understanding the way topics are organised in interaction would provide insight on the structure of dialogue beyond the sequence of utterances. However, studying this high-level structure is a complex task that we try to approach by first segmenting dialogues into smaller topically coherent sets of utterances. Understanding the interactions between these segments would then enable us to propose a model of topic organisation at a dialogue level. In this paper we work with open-domain conversations and try to reach a comparable level of accuracy as recent machine learning based topic segmentation models but with a formal approach. The features we identify as meaningful for this task help us understand better the topical structure of a conversation.