Discourse Connectives Research Articles

Some Contributions from Embodied Cognition to Psychonarratology.

Psychonarratology is a discipline that combines classic frameworks in narratology and psycholinguistics. The goal of this review article is to highlight the contributions that recent evidence from causality processing studies, conceptual metaphor theory and embodied cognition can make to Psychonarratology, in order to promote a more comprehensive study of narrative. We argue that, in order to increase its descriptive and explanatory power, Psychonarratology would benefit from a more strongly interdisciplinary approach. This approach would integrate grounded theoretical cognition and recent methods from different disciplines. With this aim, we review study evidence that highlights the fundamental link between cognition and causal connections in narratives, as well as proposals from conceptualization theories. These approaches indicate that metaphorical mappings and discourse connections play a crucial role in the establishment of narrative sequences. Moreover, we elaborate on the contributions that recent advances in research on embodied cognition, causality processing and Conceptual Metaphor Theory can make to the theoretical and methodological framework of Psychonarratology, such as promoting the design of more ecologically valid tasks, and the study of narrative production and comprehension by non-neurotypical participants.

On the influence of discourse connectives on the predictions of humans and language models.

Psycholinguistic literature has consistently shown that humans rely on a rich and organized understanding of event knowledge to predict the forthcoming linguistic input during online sentence comprehension. We, the authors, expect sentences to maintain coherence with the preceding context, making congruent sentence sequences easier to process than incongruent ones. It is widely known that discourse relations between sentences (e.g., temporal, contingency, comparison) are generally made explicit through specific particles, known as discourse connectives, (e.g., and, but, because, after). However, some relations that are easily accessible to the speakers, given their event knowledge, can also be left implicit. The goal of this paper is to investigate the importance of discourse connectives in the prediction of events in human language processing and pretrained language models, with a specific focus on concessives and contrastives, which signal to comprehenders that their event-related predictions have to be reversed. Inspired by previous work, we built a comprehensive set of story stimuli in Italian and Mandarin Chinese that differ in the plausibility and coherence of the situation being described and the presence or absence of a discourse connective. We collected plausibility judgments and reading times from native speakers for the stimuli. Moreover, we correlated the results of the experiments with the predictions given by computational modeling, using Surprisal scores obtained via Transformer-based language models. The human judgements were collected using a seven-point Likert scale and analyzed using cumulative link mixed modeling (CLMM), while the human reading times and language model surprisal scores were analyzed using linear mixed effects regression (LMER). We found that Chinese NLMs are sensitive to plausibility and connectives, although they struggle to reproduce expectation reversal effects due to a connective changing the plausibility of a given scenario; Italian results are even less aligned with human data, with no effects of either plausibility and connectives on Surprisal.

A Communication Research on Reshaping the Content Ecosystem of New Media Platforms Based on AIGC Technology

In the era of omnimedia, the rapid development of Artificial Intelligence Generated Content (AIGC) technology has profoundly and extensively influenced the news and public opinion ecosystem. This paper delves into how mainstream media utilize AIGC technology to reshape their agenda-setting capabilities within the context of intelligent communication, and exhaustively explores the specific applications of this technology on new media platforms and their resulting communication effects. Through systematic analysis, the study finds that mainstream media, empowered by intelligent communication technology, have successfully constructed a four-dimensional support system centered on discourse interpretation, discourse connectivity, discourse affinity, and discourse guidance. This system not only strengthens the discourse power and influence of mainstream media on new media platforms but also promotes the formation of a new mainstream opinion landscape, providing robust support for constructing a healthy and orderly news and public opinion ecosystem. Specifically, mainstream media utilize AIGC technology to enhance the efficiency and quality of content production, achieve precise content delivery and personalized communication, and effectively strengthen interaction and connection with audiences. Meanwhile, by reinforcing discourse interpretation and guidance, mainstream media maintain a clear orientation in the complex and ever-changing public opinion environment. The application of AIGC technology injects new vitality into the development of mainstream media in the new media era and provides essential impetus for the continuous optimization of the news and public opinion ecosystem.

Comparing the visual affordances of discrete time Markov chains and epistemic network analysis for analysing discourse connections

IntroductionResearchers in the learning sciences have been considering methods of analysing and representing group-level temporal data, particularly discourse analysis, in Computed Supported Collaborative Learning for many years.MethodsThis paper compares two methods used to analyse and represent connections in discourse, Discrete Time Markov Chains and Epistemic Network Analysis. We illustrate both methods by comparing group-level discourse using the same coded dataset of 15 high school students who engaged in group work. The groups were based on the tools they used namely the computer, iPad, or Interactive Whiteboard group. The aim here is not to advocate for a particular method but to investigate each method’s affordances.ResultsThe results indicate that both methods are relevant in evaluating the code connection within each group. In both cases, the techniques have supported the analysis of cognitive connections by representing frequent co-occurrences of concepts in a given segment of discourse.DiscussionAs the affordances of both methods vary, practitioners may consider both to gain insight into what each technique can allow them to conclude about the group dynamics and collaborative learning processes to close the loop for learners.

The Thai Discourse Treebank: Annotating and Classifying Thai Discourse Connectives

Abstract Discourse analysis is a highly applicable area of natural language processing. In English and other languages, resources for discourse-based tasks are widely available. Thai, however, has hitherto lacked such resources. We present the Thai Discourse Treebank, the first, large Thai corpus annotated in the style of the Penn Discourse Treebank. The resulting corpus has over 10,000 sentences and 18,000 instances of connectives in 33 different relations. We release the corpus alongside our list of 148 potentially polysemous discourse connectives with a total of 340 form-sense pairs and their classification criteria to facilitate future research. We also develop models for connective identification and classification tasks. Our best models achieve an F1 of 0.96 in the identification task and 0.46 on the sense classification task. Our results serve as benchmarks for future models for Thai discourse tasks.

Assessing University Students’ Competence in Translating English Discourse Connectives into Kurdish

The present study sets out to assess university students’ competence in translating English discourse connectives into Kurdish. It also attempts to study the impact of teaching on students’ translation competence for rendering these elements. Discourse connectives are one of the main subcategories of textual metadiscourse elements which play a crucial role in text organization and the creation of cohesion. Their class is mainly comprised of conjunctions and adverbials which connect ideas at intrasentential, intersentential, and textual levels. Appropriate recognition, utilization, and translation of English discourse connectives are problematic and challenging for Kurdish EFL learners and translators. Moreover, some connectors are multifunctional and polysemous in nature. This paper aims at highlighting some of these translation problems and assessing students’ competence in their rendering at the pre-test and post-test. The research design is mixed method and employs a Translation Task and a Judgement Elicitation Task as two means for measuring translation competence. These tests were carried out on 40 Third-Year students at the Department of Translation, College of Languages, University of Duhok. The study concluded that students possessed enough translation competence for making judgements on the appropriate renditions of the discourse connectives. In contrast, they failed to appropriately translate besides, although, nevertheless, yet, provided that, otherwise, unless, for, thereby, and so that in the pre-test. Furthermore, holding teaching sessions on the classifications, functions, and translation of discourse connectives had a significant impact on the acquisition of translation competence in the post-test.

AI and English language teaching: Affordances and challenges

AbstractEnglish is one of the most used languages for jobs, markets, tourism, discourse and international connectivity. However, English learners face many challenges in gaining English language skills. Extant studies show that AI has affordances to support in English language teaching and learning ELT/L. This study answers the call to examine specific challenges and affordances for using AI in ELT/L. A systematic review method was used with PRISMA principles to identify 42 studies. Findings reveal the geographical locations of studies, learner ages and years of study. Grounded coding was then used to identify affordances of the use of AI in ELT/L in the areas of speaking, writing, reading, pedagogy and self‐regulation. AI in ELT/L challenges uncovered were technology breakdowns, limited capabilities, fear and standardising language. Policymakers, funders, practitioners and educational leaders can use the information provided in this study to gain a holistic understanding of the current trend in the use of AI in ELT/L, and practical implications are provided to guide future use of AI. Practitioner notesWhat is already known about this topic English is one of the most used languages for jobs, markets, tourism, discourse and international connectivity. Empirical evidence shows that pupils can often face difficulties when learning English, with challenges such as irregularity in English spelling. AI has supported language teaching and learning with studies showing that AI can support language‐specific skills. What this paper adds Provides the scholarly community with a unique systematic review in the use of AI in ELT/L across learner levels. Identifies affordances of AI in ELT/L in speaking, writing, reading, pedagogy and self‐regulation. Identifies challenges of AI in ELT/L in technology breakdowns, limited capabilities, fear and standardising language. Provides researchers with a review of the field with identification of gaps and future research opportunities. Implications for practice and/or policy Provides practical implications from the findings for educators, policy makers and program designers. Highlights the gaps in academic knowledge as a lack in the use of AI for assessment in ELT/L.

Sense-aware connective-based indices of cohesion and their relationship to cohesion ratings of English language learners’ written production

Abstract The use of connectives has been considered important for assessing the cohesion of written texts (Crossley et al., 2019). However, existing connective-based indices have not systematically addressed two issues of ambiguity, namely, that between discourse and non-discourse use of polysemous word forms and that in terms of the specific discourse relations marked by polysemous discourse connectives (Pitler & Nenkova, 2009). This study proposes 34 sense-aware connective-based indices of cohesion that account for these issues and assesses their predictive power for cohesion ratings in comparison to 25 existing indices. Results from the analysis of 3,911 argumentative essays from the English Language Learner Insight, Proficiency and Skills Evaluation Corpus show that 23 sense-aware indices but only three existing indices correlated significantly and meaningfully with cohesion ratings. The sense-aware indices also exhibited greater predictive power for cohesion ratings than existing indices. The implications of our findings for future cohesion research are discussed.

Oral Discourse In English Of Second Language Teachers: An Analysis

This study primarily aimed to analyze the discourse competence of teachers in English as a Second Language (ESL) classrooms in the Biliran Division. Employing the mixed-methods research design, this research elicited data from the 20 secondary teachers in English as a Second Language (ESL) classrooms in Biliran Division during the school year 2021-2022. The findings reveal that the most common cohesive ties used by the English teachers were: reference (rephrasing, restatement); ellipsis (omission of word/s); substitution (the use of one); discourse markers (the use of alright, so, okay); and discourse connectives (the use of so). Moreover, the syntactic errors committed by English teachers in the Biliran Division are mostly on subject-verb agreement, use of prepositions, code-switching, pronunciation, and plurality. Based on the overall observation and analysis of the participants’ utterances, the researcher concluded that the teachers in English as a Second Language (ESL) Classrooms in Biliran Division used linguistic features as their immediate recourse in times of vocabulary loss and grammar confusion. In line with this, the school administrators may invest and strengthen the language learning development programs for ESL teachers because it is through them that the schools train and develop the minds of the next generation.

Learning dimensions of meaning: Children’s acquisition of but

Connectives such as but are critical for building coherent discourse. They also express meanings that do not fit neatly into the standard distinction between semantics and implicated pragmatics. How do children acquire them? Corpus analyses indicate that children use these words in a sophisticated way by the early pre-school years, but a small number of experimental studies also suggest that children do not understand that but has a contrastive meaning until they reach school age. In a series of eight experiments we tested children’s understanding of contrastive but compared to the causal connective so, by using a word learning paradigm (e.g., It was a warm day but/so Katy put on a pagle). When the connective so was used, we found that even 2-year-olds inferred a novel word meaning that was associated with the sentence context (a t-shirt). However, for the connective but, children did not infer a non-associated contrastive meaning (a winter coat) until age 7. Before that, even 5-year-old children reliably inferred an associated referent, indicating that they failed to correctly assign but a contrastive meaning. Five control experiments ruled out explanations for this pattern based on basic task demands, sentence processing skills or difficulty making adult-like inferences. A sixth experiment reports one particular context in which five-year-olds do interpret but contrastively. However, that same context also leads children to interpret so contrastively. We conclude that children’s sophisticated production of connectives like but and so masks a major difficulty learning their meanings. We suggest that discourse connectives incorporate a class of words whose usage is easy to mimic, but whose meanings are difficult to acquire from everyday conversations, with implications for theories of word learning and discourse processing.

Discourse-Aware Graph Networks for Textual Logical Reasoning.

Textual logical reasoning, especially question-answering (QA) tasks with logical reasoning, requires awareness of particular logical structures. The passage-level logical relations represent entailment or contradiction between propositional units (e.g., a concluding sentence). However, such structures are unexplored as current QA systems focus on entity-based relations. In this work, we propose logic structural-constraint modeling to solve the logical reasoning QA and introduce discourse-aware graph networks (DAGNs). The networks first construct logic graphs leveraging in-line discourse connectives and generic logic theories, then learn logic representations by end-to-end evolving the logic relations with an edge-reasoning mechanism and updating the graph features. This pipeline is applied to a general encoder, whose fundamental features are joined with the high-level logic features for answer prediction. Experiments on three textual logical reasoning datasets demonstrate the reasonability of the logical structures built in DAGNs and the effectiveness of the learned logic features. Moreover, zero-shot transfer results show the features' generality to unseen logical texts.

The Analysis of Discourse Markers in Science Debate

This study investigated discourse markers that occurred in the science debate by TJump and Nathan Thompson. This research aimed to determine the types of discourse markers that TJump and Nathan Thompson used and described their functions. The researcher used a descriptive-qualitative approach to answer research questions by using Schiffrin's theory (1987). The data was taken from utterances containing discourse markers from TJump and Nathan Thompson's debate that discusses globe vs. flat Earth. This research reveals that there were six types of discourse markers and 11 discourse markers, which have different functions, such as a marker of information management (oh); a marker of response (well); discourse connectives (and, but, or); markers of cause and result (so, because); markers of temporal adverbs (now, then); markers of information and participation (I know, I mean). The function of those discourse markers are a marker of information management to attract attention, a marker of response to create coherence, discourse connectives to connect more units, to mark contrasting units as option markers, a marker of cause and result as a complement, and subordinate ideas, markers of temporal adverbs are to show the relationship between time, markers of information and participation are as the transition of information state and indicate the speaker's orientation. The next researcher can use another theory to analyze discourse markers, especially discourse markers as fillers. This study focused only on the types and functions of discourse markers; the researcher can connect discourse markers to speech acts and use different theories.

Meta-learning based instance manipulation for implicit discourse relation recognition

Without discourse connectives, implicit discourse relations recognition (IDRR) remains a challenging task and has attracted increasing attention. However, most studies ignore the issues of class imbalance and data noise. To alleviate these two problems, in this paper, we propose to improve the robustness of IDRR models with two novel meta-learning based instance manipulation networks: Meta-Weight-Net (MW-Net) and Meta-Label-Net (ML-Net), which adaptively control the effect of training instances and smooth the label distributions during model training, respectively. Specifically, we use MW-Net to adaptively learn an instance weight function directly from data and integrates the instance weights into the objective function of IDRR. Meanwhile, we adopt ML-Net to dynamically refine the label distributions, leading to the smaller variance in updating gradients. We conduct experiments on the PDTB 2.0 corpus. Experimental results and in-depth analysis empirically demonstrate the effectiveness of our networks.

Usage disambiguation of Turkish discourse connectives

This paper describes a rule-based approach and a machine learning approach to disambiguate the discourse usage of Turkish connectives, which not only has single and phrasal connectives as most languages do, but also suffixal connectives that largely correspond to subordinating conjunctions in English. Since these connectives have different linguistic characteristics, two sets of linguistic rules are devised to disambiguate their discourse usage. The linguistic rules are used in the rule-based approach and employed as feature sets in the machine learning approach to test whether they influenced the decision of our algorithms. The results of both approaches are evaluated over the Turkish section of TED-Multilingual Discourse Bank and Turkish Discourse Bank 1.1, two datasets annotated in the Penn Discourse TreeBank style. The paper attests to the predictive power of the linguistic rules in disambiguating the discourse usage of both types of connectives also offering new knowledge and insights for discourse processing from the view of a morphologically rich language.

Patents, War and Peace. The Resilience of the International Patent Order Through the First World War

By the end of the 19th century, an international order had emerged for patents, allowing business actors to use patents in many countries concurrently, and thus supporting a new phase in the development of industrial capitalism. Centred on Europe in spite of a universalistic discourse and important transatlantic connections, this international order was shaken by the outbreak of the First World War. Indeed, patents were impacted by provisions of economic warfare. These measures, although similar in their principle in all warring countries, were however executed to a varying degree, depending on the extent of economic dependence or fear of retaliation. These differences also manifested during the negotiation of the Treaty of Versailles, the analysis of which contradicts the idea of a general expropriation of German patents. Thus escaping disintegration, the international patent order was afterwards progressively restored on its pre-1914 foundations.

A constructional account of the development of the Chinese stance discourse markerběnlái

AbstractThis paper examines the emergence of the Chinese adverbběnláiand its further development into a stance discourse marker (sdm). Most previous studies consider the epistemic modal adverbběnlái(‘indeed, truly’) to have originated from the temporal adverbběnlái(‘originally’). Based on the framework of constructionalisation (Traugott and Trousdale 2013) andVan de Velde et al.’s (2013)idea of multiple sources, I argue that the temporal adverb might not have been the only source for the modalběnláito occur. Furthermore, the modal adverb changed into ansdm, thereby signaling the speaker’s subjective evaluation while linking the coherence of the discourse. I argue that the change was enabled by both subjectification and analogisation. The exemplars to whichběnláiwas by hypothesis analogised include the extant discourse connectives (e.g.,kěshì) and the extant commentary pragmatic markers (e.g.,kěliánandsuǒxìng) (Fraser 2009).

Discourse connectives and their arguments: an experiment on anaphoricity in German

Abstract Adverbial connectives like therefore, which link a preceding ‘external’ to an ‘internal’ argument, can be regarded as anaphoric: The external argument is selected by an interpretation process akin to that of an event anaphor, and intervening material can appear between both arguments. We report on a crowdsourcing experiment on the German connectives trotzdem and dennoch that studies factors that lead readers to assume such long-distance arguments: semantic plausibility of intervening material, ‘subjective’ versus ‘objective’ content, and the presence of an anaphoric morpheme in the connective. We find that the type and content of the intervening material play an important role in argument choice.

The Signaling of Continuative and Contrastive Discourse Relations in English Argumentative Discourse: Corpus-Based and Experimental Perspectives

This paper examines the linguistic realization of continuative and contrastive discourse relations (DRs) in English argumentative discourse, focusing on both discourse as product and discourse as process. Based on a corpus of 25 argumentative texts with corresponding experimental data stemming from an editing-based task, this study aims to answer the questions of (i) how continuative and contrastive DRs are linguistically realized (i.e., encoded in coherence strands or additionally signaled) in English argumentative discourse, and (ii) how their signaling unfolds during the real-time process of discourse editing. The quantitative analysis of the edited texts shows that contrastive DRs are signaled throughout (e.g., through contrastive discourse connectives), whereas continuative DRs are more frequently encoded only. In addition, the qualitative analysis of the experimental data (i.e., keystroke logging data and corresponding metadata) reveals that while the signaling of contrastive DRs is made manifest right from the start of the editing process across the data, continuative DRs tend to be signaled only at a later stage. This can be explained by the semantic overlap between different types of continuative DRs, which requires that the participants (re-)negotiate the exact nature (and thus the appropriate linguistic signal) of a particular continuative DR more frequently than contrastive DRs, which are clearly interpreted and thus readily signaled as contrastive throughout.

Individual Variations in the Mastery of Discourse Connectives from Teenage Years to Adulthood

AbstractMany connectives, such as therefore and however, are used very frequently in the written modality. Their acquisition thus represents an important milestone in developing written language competences. In this article, we assess the development of competence with such connectives by native French speakers in a sentence‐level insertion task (N = 307, aged 12 to 64) and a text‐level insertion task (N = 172, aged 13 to 71). Our results indicate that, despite a general progression in the level of competence with age, the academic level of participants is a strong predictor of competence within each age group, even during adulthood. In addition, from the age of 12, competence is related to the frequency of connectives in naturalistic data, with frequent connectives systematically mastered better than less frequent ones. Finally, in all age groups, the use and understanding of connectives is more challenging when sentences to complete are embedded within a richer context than when presented alone.

How Robust Is Discourse Processing for Native Readers? The Role of Connectives and the Coherence Relations They Convey.

While corpus studies have shown that discourse connectives that convey the same coherence relation can display subtle differences, research on online discourse processing has only focused on a rather limited set of connectives. Yet, different connectives – for example, rare or polyfunctional ones – might elicit different reading patterns. In order to explore this assumption, we test the robustness of discourse processing for French native speakers by measuring the way they process causal and concessive sentences that are conveyed by either an appropriate or inappropriate connective. Throughout three experiments, we change important characteristics of the connectives: we first test frequently used connectives (Experiment 1), secondly less frequent ones (Experiment 2), and finally less frequent connectives that are polyfunctional and for which different functions clearly compete (Experiment 3). Our results show that the processing for incoherent items was affected for all connectives, however readers showed altered reading fluency when infrequent connectives were used. We conclude that discourse processing is quite robust and that readers are able to insert meaning conveyed by rare connectives while still showing the highest reading ease with frequent connectives.