Recent LLM benchmarks have tested models on a range of phenomena, but are still focused primarily on natural language understanding for extraction of explicit information, such as QA or summarization, with responses often targeting information from individual sentences. We are still lacking more challenging, and importantly also multilingual, benchmarks focusing on implicit information and pragmatic inferences across larger documents in the context of discourse tracking: integrating and aggregating information across sentences, paragraphs and multiple speaker utterances. To this end, we present DiscoTrack, an LLM benchmark targeting a range of tasks across 12 languages and four levels of discourse understanding: salience recognition, entity tracking, discourse relations and bridging inference. Our evaluation shows that these tasks remain challenging, even for state-of-the-art models.

The role of world knowledge has been particularly crucial to predict the discourse connective that marks the discourse relation between two arguments, with language models (LMs) being generally successful at this task. We flip this premise in our work, and instead study the inverse problem of understanding whether discourse connectives can inform LMs about the world. To this end, we present WUGNECTIVES, a dataset of 8,880 stimuli that evaluates LMs' inferences about novel entities in contexts where connectives link the entities to particular attributes. On investigating 17 different LMs at various scales, and training regimens, we found that tuning an LM to show reasoning behavior yields noteworthy improvements on most connectives. At the same time, there was a large variation in LMs' overall performance across connective type, with all models systematically struggling on connectives that express a concessive meaning. Our findings pave the way for more nuanced investigations into the functional role of language cues as captured by LMs. We release WUGNECTIVES at https://github.com/sheffwb/wugnectives.

Research in linguistics shows that non-verbal cues, such as gestures, play a crucial role in spoken discourse. For example, speakers perform hand gestures to indicate topic shifts, helping listeners identify transitions in discourse. In this work, we investigate whether the joint modeling of gestures using human motion sequences and language can improve spoken discourse modeling in language models. To integrate gestures into language models, we first encode 3D human motion sequences into discrete gesture tokens using a VQ-VAE. These gesture token embeddings are then aligned with text embeddings through feature alignment, mapping them into the text embedding space. To evaluate the gesture-aligned language model on spoken discourse, we construct text infilling tasks targeting three key discourse cues grounded in linguistic research: discourse connectives, stance markers, and quantifiers. Results show that incorporating gestures enhances marker prediction accuracy across the three tasks, highlighting the complementary information that gestures can offer in modeling spoken discourse. We view this work as an initial step toward leveraging non-verbal cues to advance spoken language modeling in language models.

Social media users drive the spread of misinformation online by sharing posts that include erroneous information or commenting on controversial topics with unsubstantiated arguments often in earnest. Work on echo chambers has suggested that users' perspectives are reinforced through repeated interactions with like-minded peers, promoted by homophily and bias in information diffusion. Building on long-standing interest in the social bases of language and linguistic underpinnings of social behavior, this work explores how conversations around misinformation are mediated through language use. We compare a number of linguistic measures, e.g., in-/out-group cues, readability, and discourse connectives, within and across topics of conversation and user communities. Our findings reveal increased presence of group identity signals and processing fluency within echo chambers during discussions of misinformation. We discuss the specific character of these broader trends across topics and examine contextual influences.

In this work, we introduce a lightweight discourse connective detection system. Employing gradient boosting trained on straightforward, low-complexity features, this proposed approach sidesteps the computational demands of the current approaches that rely on deep neural networks. Considering its simplicity, our approach achieves competitive results while offering significant gains in terms of time even on CPU. Furthermore, the stable performance across two unrelated languages suggests the robustness of our system in the multilingual scenario. The model is designed to support the annotation of discourse relations, particularly in scenarios with limited resources, while minimizing performance loss.

Abduction has long been seen as crucial for narrative comprehension and reasoning about everyday situations. The abductive natural language inference ($\alpha$NLI) task has been proposed, and this narrative text-based task aims to infer the most plausible hypothesis from the candidates given two observations. However, the inter-sentential coherence and the model consistency have not been well exploited in the previous works on this task. In this work, we propose a prompt tuning model $\alpha$-PACE, which takes self-consistency and inter-sentential coherence into consideration. Besides, we propose a general self-consistent framework that considers various narrative sequences (e.g., linear narrative and reverse chronology) for guiding the pre-trained language model in understanding the narrative context of input. We conduct extensive experiments and thorough ablation studies to illustrate the necessity and effectiveness of $\alpha$-PACE. The performance of our method shows significant improvement against extensive competitive baselines.

Implicit discourse relation classification is a challenging task due to the absence of discourse connectives. To overcome this issue, we design an end-to-end neural model to explicitly generate discourse connectives for the task, inspired by the annotation process of PDTB. Specifically, our model jointly learns to generate discourse connectives between arguments and predict discourse relations based on the arguments and the generated connectives. To prevent our relation classifier from being misled by poor connectives generated at the early stage of training while alleviating the discrepancy between training and inference, we adopt Scheduled Sampling to the joint learning. We evaluate our method on three benchmarks, PDTB 2.0, PDTB 3.0, and PCC. Results show that our joint model significantly outperforms various baselines on three datasets, demonstrating its superiority for the task.

Discourse relations are typically modeled as a discrete class that characterizes the relation between segments of text (e.g. causal explanations, expansions). However, such predefined discrete classes limits the universe of potential relationships and their nuanced differences. Analogous to contextual word embeddings, we propose representing discourse relations as points in high dimensional continuous space. However, unlike words, discourse relations often have no surface form (relations are between two segments, often with no word or phrase in that gap) which presents a challenge for existing embedding techniques. We present a novel method for automatically creating discourse relation embeddings (DiscRE), addressing the embedding challenge through a weakly supervised, multitask approach to learn diverse and nuanced relations between discourse segments in social media. Results show DiscRE can: (1) obtain the best performance on Twitter discourse relation classification task (macro F1=0.76) (2) improve the state of the art in social media causality prediction (from F1=.79 to .81), (3) perform beyond modern sentence and contextual word embeddings at traditional discourse relation classification, and (4) capture novel nuanced relations (e.g. relations semantically at the intersection of causal explanations and counterfactuals).

We present DiscoSense, a benchmark for commonsense reasoning via understanding a wide variety of discourse connectives. We generate compelling distractors in DiscoSense using Conditional Adversarial Filtering, an extension of Adversarial Filtering that employs conditional generation. We show that state-of-the-art pre-trained language models struggle to perform well on DiscoSense, which makes this dataset ideal for evaluating next-generation commonsense reasoning systems.

We describe Turkish Discourse Bank 1.2, the latest version of a discourse corpus annotated for explicitly or implicitly conveyed discourse relations, their constitutive units, and senses in the Penn Discourse Treebank style. We present an evaluation of the recently added tokens and examine three commonly occurring dependency patterns that hold among the constitutive units of a pair of adjacent discourse relations, namely, shared arguments, full embedding and partial containment of a discourse relation. We present three major findings: (a) implicitly conveyed relations occur more often than explicitly conveyed relations in the data; (b) it is much more common for two adjacent implicit discourse relations to share an argument than for two adjacent explicit relations to do so; (c) both full embedding and partial containment of discourse relations are pervasive in the corpus, which can be partly due to subordinator connectives whose preposed subordinate clause tends to be selected together with the matrix clause rather than being selected alone. Finally, we briefly discuss the implications of our findings for Turkish discourse parsing.