id
stringlengths
20
52
title
stringlengths
3
459
abstract
stringlengths
0
12.3k
classification_labels
sequence
numerical_classification_labels
sequence
SCOPUS_ID:85111160496
“accipiendo vel compilando ab aliis”: De vulgari eloquentia 1.1. Reading Dante with Dante: A contribution to Dante’s theory of language
This article analyzes Dante’s theory of language and considers at first a few fragments of Dante’s Latin treatise on the vernacular, reading them in light of their ancient-medieval contexts. This reading allows part-modification of the critical discourse about Dante’s theory of language. The article argues that Dante’s discussion did not start in the De vulgari eloquentia, as is commonly assumed, but was at first introduced in the Vita nuova. Recent studies show that the theme of laude in the Vita nuova includes a linguistic theory and a discourse on the deep structures of language. Focussing on specific words, considering them in light of the ancient-medieval background, the article organizes a transverse reading that considers layers of Dante’s discourse on language from the Vita nuova to the Commedia not yet explored and evaluated.
[ "Linguistics & Cognitive NLP", "Linguistic Theories" ]
[ 48, 57 ]
SCOPUS_ID:85090896547
“no, professor, that is not true”: First attempts at introducing translanguaging to pre–service teachers
This chapter considers the complexities of conceptualizing and introducing translanguaging to pre-service teachers by analyzing the first author’s experiences as a teacher educator in her initial effort at discussing translanguaging and its pedagogical implications in a second language learning course. Discussions about translanguaging were marked by respectful but contentious debate, in which students differed strongly from their teacher in their responses and opinions related to translanguaging and its relationship to code-switching. This study’s purpose, conducted in the spirit of teacher inquiry, is to examine the reflections of both the instructor and her students in seeking to understand and interpret the literature on translanguaging. We found that this process can pose significant challenges for teacher educators and their students and raises important questions about the relationship between scholarly and practitioner communities in the language education field.
[ "Code-Switching", "Multilinguality" ]
[ 7, 0 ]
SCOPUS_ID:85092150442
“oh my god! buy it!” a multimodal discourse analysis of the discursive strategies used by chinese ecommerce live-streamer austin li
Ecommerce livestreaming, also known as live commerce or social commerce, has taken off over the past two years in East Asia and is showing the tendency of going global. Intrigued by the phenomenal success of ecommerce livestream, we concentrate on analyzing the most prominent and illustrative example of Chinese ecommerce live-streamer Austin Li. Through this individual case study, we aim to investigate discursive strategies employed in ecommerce livestreaming and reveal resources specific to this new media genre. Guided by multimodal discourse analysis, our research first accommodates the socio-economic context of Li’s success to warrant social situatedness in interpreting data. After that we move into analyzing his discourse employed in livestreaming. Research findings suggest that in attention economy, Li strategically utilizes his male gender as a resource in trying on lipsticks for female customers. His discourse in multiple modes serves to build consumer trust and propagate products. An in-depth analysis of his discursive strategies indicates that, ecommerce livestreaming as a new form of advertising not only shares commonalities with traditional advertisement discourse but also embodies affordances that are specific to livestreaming platforms. To be more specific, livestreaming is featured with delimitation of time, real-time interactivity, and video-aided communication. These affordances enable Li to adopt more interactive and personalized persuasive discourse than conventional advertisements.
[ "Discourse & Pragmatics", "Semantic Text Processing", "Multimodality" ]
[ 71, 72, 74 ]
https://aclanthology.org//2022.wassa-1.4/
“splink” is happy and “phrouth” is scary: Emotion Intensity Analysis for Nonsense Words
People associate affective meanings to words - “death” is scary and sad while “party” is connotated with surprise and joy. This raises the question if the association is purely a product of the learned affective imports inherent to semantic meanings, or is also an effect of other features of words, e.g., morphological and phonological patterns. We approach this question with an annotation-based analysis leveraging nonsense words. Specifically, we conduct a best-worst scaling crowdsourcing study in which participants assign intensity scores for joy, sadness, anger, disgust, fear, and surprise to 272 non-sense words and, for comparison of the results to previous work, to 68 real words. Based on this resource, we develop character-level and phonology-based intensity regressors. We evaluate them on both nonsense words and real words (making use of the NRC emotion intensity lexicon of 7493 words), across six emotion categories. The analysis of our data reveals that some phonetic patterns show clear differences between emotion intensities. For instance, s as a first phoneme contributes to joy, sh to surprise, p as last phoneme more to disgust than to anger and fear. In the modelling experiments, a regressor trained on real words from the NRC emotion intensity lexicon shows a higher performance (r = 0.17) than regressors that aim at learning the emotion connotation purely from nonsense words. We conclude that humans do associate affective meaning to words based on surface patterns, but also based on similarities to existing words (“juy” to “joy”, or “flike” to “like”).
[ "Emotion Analysis", "Sentiment Analysis" ]
[ 61, 78 ]
SCOPUS_ID:85079535422
“there is no there there”: Space deictics, verb tense, and nostalgia at a family literacy class
Drawing from a yearlong ethnographic study, this paper examines spatial deictics and verb tense use in conversation and instructional activities in a family literacy class at a Bay Area Public Library. More specifically, we employ discourse analytic tools to document how spatiotemporal coordinates are never solely product of cognitive calculation but always also entangled with emotions. In addition to enacting their referentiality with respect to places and moments in time, spatial and temporal indexical terms served to position participants within a discursive paradigm of migration, a chronotope imbued by nostalgia. We discuss how such discursive positioning constructs students’ identities as diasporic, their being in the here-and-now predicated on their being from elsewhere, and their being here-and-there simultaneous and coextensive.
[ "Discourse & Pragmatics", "Semantic Text Processing" ]
[ 71, 72 ]
SCOPUS_ID:84918832305
“well, Hang on, They’re actually much better than that!”: Disrupting dominant discourses of deficit about English language learners in senior high school English
This paper explores how four English teachers position their English language learners for critical literacy within senior high school curriculum in Queensland, Australia. Such learners are often positioned, even by their teachers, within a broader “deficit discourse” that claims they are inherently lacking the requisite knowledge and skills to engage with intransigent school curricula. As such, English language learners’ identity formation is often constrained by deficit views that can ultimately see limited kinds of literacy teaching offered to them. Using Fairclough’s (2003) critical discourse analysis method, analysis of 16 interviews with the teachers was conducted as part of a larger, critical instrumental case study in two state high schools during 2010. Five competing discourses were identified: deficit as lack; deficit as need; learner “difference” as a resource; conceptual capacity for critical literacy; and linguistic, cultural and conceptual difficulty with critical literacy. While a deficit view is present, counter-hegemonic discourses also exist in their talk. The combination of discourses challenges monolithic deficit views of English language learners, and opens up generative discursive territory to position English language learners in ways other than “problematic”. This has important implications for how teachers view and teach English language learners and their capacity for critical literacy work in senior high school classrooms.
[ "Discourse & Pragmatics", "Semantic Text Processing" ]
[ 71, 72 ]
SCOPUS_ID:85147556475
“¿Qué pinga es esto?” Dysphemistic uses in Cuban Spanish
The research aims to examine the path that the lexeme pinga has had in the Spanish language, specifically in the Cuban variant of Spanish, to identify its current uses in this variety, and to describe the semantic, pragmatic, and discursive procedures that are manifested in them. It is based on the most recent linguistic theories about the taboo that define euphemistic and dysphemistic uses as cognitive processes of conceptualization of a forbidden reality. From a methodological point of view, the word pinga is traced in the main academic and Cuban dictionaries and a corpus was created from the entries for Cuba in CREA and CORPES XXI academic corpora, as well as the Cuban news portal CiberCuba. The search in dictionaries made possible the conclusion that, through the use of metaphor, pinga transferred its meaning to penis, which has been the only meaning recorded in Cuban dictionaries since the 19th century. In current Cuban Spanish, the word can be considered polysemic. Along with the meaning of the male sexual organ, there are others that show a process of loss of referential meaning, of resemantization and desemantization, common in euphemisms and dysphemisms. The uses that are identified correspond to the most frequent dysphemistic uses: insult, interjection, and intensification.
[ "Linguistics & Cognitive NLP", "Linguistic Theories" ]
[ 48, 57 ]
SCOPUS_ID:85133577129
„Garbage in, Garbage out”. The Impact of Coders’ Quality on the Neural Network Classifying Text on Social Media
One of the critical decisions when manually coding text data is whether to verify the coders’ work. In the case of supervised models, this leads to a significant dilemma: is it better to provide the model with a large number of cases on which it will learn at the expense of verifying the correctness of the data, or whether it is better to code each case n-times, which will allow to compare the codes and check their correctness but at the same time will reduce the training dataset by n-fold. Such a decision not only affect the final results of the classifier. From the researchers’ point of view, it is also crucial because, realistically assuming that research has limited funding, it cannot be undone. The study uses a simulation approach and provides conclusions and recommendations based on 100,000 unique and hand-coded tweets.
[ "Information Retrieval", "Text Classification", "Information Extraction & Text Mining" ]
[ 24, 36, 3 ]
SCOPUS_ID:85148623860
„MIS KEELES MA RÄÄGIN, I DON’T KNOW“. ON THE USE OF ENGLISH BY ESTONIAN YOUTUBERS
This article focuses on English use in the example of Estonian-speaking YouTubers. Altogether, we analysed videos from eight content creators, each well-known among high-school-aged viewers who post regular videos in Estonian. The dataset consists of videos (or video excerpts) in which we look into the proportional share of English words or phrases and explore potential functions of code-switching. The results show that while all eight YouTubers use English in multiple videos, the usage frequencies differ significantly and reflect individual differences. English emerged in platform-specific contexts where the words were directly related to content creation (26% of all code-switching cases). Occasionally, the speakers referred to English pop culture phenomena (16%), expressed emotions (12%) and used loanwords or other (embedded) elements (6%). For numerous cases (23%), it was hard to determine why they preferred using an English word or phrase instead of its Estonian equivalent.
[ "Code-Switching", "Visual Data in NLP", "Multimodality", "Multilinguality" ]
[ 7, 20, 74, 0 ]
SCOPUS_ID:85045878911
„Reading between the lines of law“. About different functions of the siete partidas of king alfons X. of Castile
The Siete Partidas, the famous law code, of Alfons X. of Castile, written in the 13th century offers us a great variety of topics which have to be investigated more detailed. The central question is, which functions has the text besides being a law code? The encyclopedic character of this work, provides an insight into daily life of medieval Castile and mirrors the historical frame in which it was written. One of the central ideas of the Siete Partidas was to unify the existing „laws“ in the Kingdom of Castile. But, unification also means a centralization of the power which was embodied by the king. The fact, that this was not completely accepted by other powerful people of that time made it necessary for Alfons X. to legitimate his own power. This article highlights different functions of the Siete Partidas of King Alfons X. of Castile based on a discourse analysis and a hermeneutical interpretation.
[ "Discourse & Pragmatics", "Semantic Text Processing" ]
[ 71, 72 ]
SCOPUS_ID:84941401628
„Sentence patterns‟ in the light of language theories and classroom needs
[ "Linguistics & Cognitive NLP", "Linguistic Theories" ]
[ 48, 57 ]
SCOPUS_ID:85069700929
„Sie scheint auch mehr zu donnern und zu blitzen, als zu reden.“ Zur Meteorologisierung der Sprache im Drama der 1770er Jahre
This article offers an account of wettern (weathering) in German drama of the 1770s, with focus on works by Goethe, Klinger, Lenz, and Herder. Extant scholarship on this period tends to stress the religious and rhetorical origins, as well as the allegorical function, of meteorological imagery. Against these readings, this article argues that instances of wettern must be read literally, insofar as the works under examination are informed by then-emerging meteorological science, whose hallmark is the synthesis of formerly separate phenomena (e.g., lightning, clouds, sunlight, storms) into a continuous process with a unified internal dynamic. On this account, the dynamization of weather in modern meteorology sparked a dynamization of rhetorical formula in German dramatic language, whereby wettern became prototypical of the dynamic speech that others have observed to appear in 1770s drama. To reveal the programmatic significance of wettern across these works, attention is given, first of all, to metafictional moments that assert a necessary relation between weather and drama as a medium, and, second, to the overlooked role of weather in Herder’s early theory of language, which was a formative influence on the authors considered here. In both cases, Lessing’s disparaging view of scientific and dramatic representations of meteorological phenomena is offered, by way of contrast, to show how the poetics of weather developed in and through wettern departed from the status quo of aesthetic theory.
[ "Linguistics & Cognitive NLP", "Linguistic Theories" ]
[ 48, 57 ]
SCOPUS_ID:85077663003
„The only image in our heart.“Word and picture in M. Luther’s theological thinking
The essay profiles Luther’s „Theology of the image“within his „Theology of the word“, integrated in the visualized context of the „iconic turn“. The relationship of „word and image“is explicated anthropologically, christologically and soteriologically . Christ, the logos and the only image, as Luther describes especially in his pastoral scriptures and sermons, is the foundation and the perspective of the Christian „vision-school of faith“, which embraces the „heart“and the senses . Modern theories of language and image deepen Luther’s theology of word and image, which is grounded in the word of the triune God and which is concentrated in the justification of sinners by grace, as it is promised in the witness of the gospel and in the care of souls to people, who look for consolation and help . In addition the theology of word and image of the existential thinking of the Reformer also opens an ecumenical conversation .
[ "Visual Data in NLP", "Multimodality", "Linguistics & Cognitive NLP", "Linguistic Theories" ]
[ 20, 74, 48, 57 ]
SCOPUS_ID:84902147873
•” Media: Text Based CMC Tool Which Touch off Informal Communication
This paper describes about the media which touch off informal text communication by black spot stimulus. A black spot stimulus is shown as a background image of the text field in advance of text input. The purpose of black spot stimulus presentation is to touch off conceiving topic based on the cognitive process model of creative thinking. The prototype system based on this media concept was developed, and the communication experiment was conducted. It turned out as a result of the experiment that these media touched off the idea about making topic strongly in the text communication. Moreover, it was implied that the topic touched off deepens the mutual understanding between speakers. © 2010, The Institute of Image Electronics Engineers of Japan. All rights reserved.
[ "Visual Data in NLP", "Multimodality" ]
[ 20, 74 ]
SCOPUS_ID:85055629876
ℓ<inf>0</inf>-based sparse canonical correlation analysis with application to cross-language document retrieval
Most of existing sparse CCA algorithms compute sparse weight vectors by minimizing the ℓ1 norm, which imposes essential difficulty for the analysis of the solution. Different from existing ones, this paper develops a novel sparse CCA algorithm by ℓ0 penalty. The resulting ℓ0 minimization problem is solved by means of residual, which has one merit that no regularization parameter or shrinkage parameter needs to be tuned. We also provide consistency analysis of the proposed method using the concept of Restricted Isometry Property (RIP) condition, while no theoretical guarantee was given for most of existing sparse CCA methods. Sparsity bound of the CCA solutions is also studied. Experimental results on both simulated dataset and real-world datasets in cross-language document retrieval task demonstrate the effectiveness and competitiveness of the proposed algorithm, when compared with several state-of-the-art sparse CCA methods.
[ "Document Retrieval", "Information Retrieval" ]
[ 56, 24 ]
SCOPUS_ID:85052372839
ℓ<inf>1</inf> Regularization of word embeddings for multi-word expression identification
In this paper we compare the effects of applying various state-of-the-art word representation strategies in the task of multi-word expression (MWE) identification. In particular, we analyze the strengths and weaknesses of the usage of `1-regularized sparse word embeddings for identifying MWEs. Our earlier study demonstrated the effectiveness of regularized word embeddings in other sequence labeling tasks, i.e. part-of-speech tagging and named entity recognition, but it has not yet been rigorously evaluated for the identification of MWEs yet.
[ "Semantic Text Processing", "Representation Learning" ]
[ 72, 12 ]
SCOPUS_ID:84862996515
ℳ-adhesive transformation systems with nested application conditions. part 2: Embedding, critical pairs and local confluence
Graph transformation systems have been studied extensively and applied to several areas of computer science like formal language theory, the modeling of databases, concurrent or distributed systems, and visual, logical, and functional programming. In most kinds of applications it is necessary to have the possibility of restricting the applicability of rules. This is usually done by means of application conditions. In this paper, we continue the work of extending the fundamental theory of graph transformation to the case where rules may use arbitrary (nested) application conditions. More precisely, we generalize the Embedding theorem, and we study how local confluence can be checked in this context. In particular, we define a new notion of critical pair which allows us to formulate and prove a Local Confluence Theorem for the general case of rules with nested application conditions. All our results are presented, not for a specific class of graphs, but for any arbitrary ℳ-adhesive category, which means that our results apply to most kinds of graphical structures. We demonstrate our theory on the modeling of an elevator control by a typed graph transformation system with positive and negative application conditions. © 2012 IOS Press.
[ "Semantic Text Processing", "Linguistic Theories", "Structured Data in NLP", "Representation Learning", "Linguistics & Cognitive NLP", "Multimodality" ]
[ 72, 57, 50, 12, 48, 74 ]
SCOPUS_ID:85115826050
⇜Do users need human-like conversational agents?” - Exploring conversational system design using framework of human needs
The fascinating story of human evolution can be attributed to our ability to speak, write, and communicate complex thoughts. When researchers envision a perfect, artificially intelligent conversational system, they want the system to be human-like. In other words, the system should converse with the same intellect and cognition as humans. Now, the question which we need to ask is if we need a human-like conversational system? Before we engage in the complex endeavor of implementing human-like characteristics, we should debate if the pursuit of such a system is logical and ethical. We analyze some of the system-level characteristics and discuss their merits and potential of harm. We review some of the latest work on conversational systems to understand how design features are evolving for Conversational Agents. Additionally, we look into the framework of human needs to assess how the system should assign relative importance to user requests, and prioritize user tasks. We draw on the peer work in human-computer interaction, sentiment analysis, and human psychology to provide insights into how future conversational agents should be designed for better user satisfaction.
[ "Natural Language Interfaces", "Dialogue Systems & Conversational Agents" ]
[ 11, 38 ]
SCOPUS_ID:85134820297
∗SEM 2016 - 5th Joint Conference on Lexical and Computational Semantics, Proceedings
The proceedings contain 27 papers. The topics discussed include: quantificational features in distributional word representations; automatic identification of aspectual classes across verbal readings; metaphor as a medium for emotion: an empirical study; driving inversion transduction grammar induction with semantic evaluation; how factuality determines sentiment inferences; sense embedding learning for word sense induction; improving zero-shot-learning for German particle verbs by using training-space restrictions and local scaling; and when hyperparameters help: beneficial parameter combinations in distributional semantic models.
[ "Text Error Correction", "Syntactic Text Processing" ]
[ 26, 15 ]
SCOPUS_ID:85144363584
♠ SPADE: A Big Five-Mturk Dataset of Argumentative Speech Enriched with Socio-Demographics for Personality Detection
In recent years, there has been increasing interest in automatic personality detection based on language. Progress in this area is highly contingent upon the availability of datasets and benchmark corpora. However, publicly available datasets for modeling and predicting personality traits are still scarce. While recent efforts to create such datasets from social media (Twitter, Reddit) are to be applauded, they often do not include continuous and contextualized language use. In this paper, we introduce ♠ SPADE, the first dataset with continuous samples of argumentative speech labeled with the Big Five personality traits and enriched with socio-demographic data (age, gender, education level, language background). We provide benchmark models for this dataset to facilitate further research and conduct extensive experiments. Our models leverage 436 (psycho)linguistic features extracted from transcribed speech and speaker-level metainformation with transformers. We conduct feature ablation experiments to investigate which types of features contribute to the prediction of individual personality traits.
[ "Linguistics & Cognitive NLP", "Speech & Audio in NLP", "Psycholinguistics", "Multimodality" ]
[ 48, 70, 77, 74 ]
https://aclanthology.org//2020.webnlg-1.10/
𝒫2: A Plan-and-Pretrain Approach for Knowledge Graph-to-Text Generation
Text verbalization of knowledge graphs is an important problem with wide application to natural language generation (NLG) systems. It is challenging because the generated text not only needs to be grammatically correct (fluency), but also has to contain the given structured knowledge input (relevance) and meet some other criteria. We develop a plan-and-pretrain approach, 𝒫2, which consists of a relational graph convolutional network (RGCN) planner and the pretrained sequence-tosequence (Seq2Seq) model T5. Specifically, the R-GCN planner first generates an order of the knowledge graph triplets, corresponding to the order that they will be mentioned in text, and then T5 produces the surface realization of the given plan. In the WebNLG+ 2020 Challenge, our submission ranked in 1st place on all automatic and human evaluation criteria of the English RDF-to-text generation task.
[ "Language Models", "Semantic Text Processing", "Structured Data in NLP", "Knowledge Representation", "Text Generation", "Data-to-Text Generation", "Multimodality" ]
[ 52, 72, 50, 18, 47, 16, 74 ]