Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18265 papers
Natural Language Toolkit (NLTK) is a comprehensive Python library designed to facilitate the exploration, analysis, and processing of human language data. With its extensive collection of tools, NLTK provides researchers, developers, and educators with a powerful platform for tasks ranging from basic text processing to advanced natural language understanding and machine learning. The toolkit includes modules for tokenization, stemming, lemmatization, part-of-speech tagging, named entity recognition, syntactic parsing, semantic analysis, and more. Furthermore, NLTK offers access to numerous linguistic resources such as corpora, lexicons, and treebanks, making it an invaluable resource for both learning and research in the field of natural language processing (NLP). NLTK serves as an indispensable tool for unlocking the complexities of human language
In everyday life, music is increasingly being listened to through headphones and mobile devices in public situations. While a large body of research has demonstrated that music may influence the emotional states of listeners and affect multimodal perceptions i.e. in films, less is known about the music’s impact on environments and the interpretation of social situations. We conducted an online experiment to investigate the influence of music on evaluations, considering individuals’ emotional states (emotion congruence) and group perception. Participants were randomly assigned to one of three experimental conditions (music with positive valence and high arousal, music with negative valence and low arousal, and no music) while viewing images of two different social group types that varied in perceived group characteristics (group members being familiar or unfamiliar with each other). Images were rated on four bipolar scales measuring affective quality and cognitive evaluation of social situations. Results show that individuals who listened to negative music provided lower valence ratings and also judgded social environments lower in terms of pleasantness and cheerfulness (affective) than individuals in the other experimental conditions. In contrast, ratings of crowdedness and familiarity (cognitive) did not differ between experimental conditions. The effect of music on affective evaluations was shaped by social group types, such that participants were more influenced by music when viewing intimacy groups (e.g., friends) than when viewing transitory groups (e.g., strangers). Overall, our results support the assumption of mood congruency for affective evaluations and emphasize the need to consider social information when studying the influence of music on the perception of environments.
Direct dependency parsing of the speech signal -- as opposed to parsing speech transcriptions -- has recently been proposed as a task (Pupier et al. 2022), as a way of incorporating prosodic information in the parsing system and bypassing the limitations of a pipeline approach that would consist of using first an Automatic Speech Recognition (ASR) system and then a syntactic parser. In this article, we report on a set of experiments aiming at assessing the performance of two parsing paradigms (graph-based parsing and sequence labeling based parsing) on speech parsing. We perform this evaluation on a large treebank of spoken French, featuring realistic spontaneous conversations. Our findings show that (i) the graph based approach obtain better results across the board (ii) parsing directly from speech outperforms a pipeline approach, despite having 30% fewer parameters.
Abstract Human creativity originates from brain cortical networks that are specialized in idea generation, processing, and evaluation. The concurrent verbalization of our inner thoughts during the execution of a design task enables the use of dynamic semantic networks as a tool for investigating, evaluating, and monitoring creative thought. The primary advantage of using lexical databases such as WordNet for reproducible information-theoretic quantification of convergence or divergence of design ideas in creative problem solving is the simultaneous handling of both words and meanings, which enables interpretation of the constructed dynamic semantic networks in terms of underlying functionally active brain cortical regions involved in concept comprehension and production. In this study, the quantitative dynamics of semantic measures computed with a moving time window is investigated empirically in the DTRS10 dataset with design review conversations and detected divergent thinking is shown to predict success of design ideas. Thus, dynamic semantic networks present an opportunity for real-time computer-assisted detection of critical events during creative problem solving, with the goal of employing this knowledge to artificially augment human creativity.
Social decision-making is known to be influenced by predictive emotions or the perceived reciprocity of partners. However, the connection between emotion, decision-making, and contextual reciprocity remains less understood. Moreover, arguments suggest that emotional experiences within a social context can be better conceptualised as prosocial rather than basic emotions, necessitating the inclusion of two social dimensions: focus, the degree of an emotion's relevance to oneself or others, and dominance, the degree to which one feels in control of an emotion. For better representation, these dimensions should be considered alongside the interoceptive dimensions of valence and arousal. In an ultimatum game involving fair, moderate, and unfair offers, this online study measured the emotions of 476 participants using a multidimensional affective rating scale. Using unsupervised classification algorithms, we identified individual differences in decisions and emotional experiences. Certain individuals exhibited consistent levels of acceptance behaviours and emotions, while reciprocal individuals' acceptance behaviours and emotions followed external reward value structures. Furthermore, individuals with distinct emotional responses to partners exhibited unique economic responses to their emotions, with only the reciprocal group exhibiting sensitivity to dominance prediction errors. The study illustrates a context-specific model capable of subtyping populations engaged in social interaction and exhibiting heterogeneous mental states.
This paper identifies a micro-cue correlating to verb second word order (V2) in two closely related Medieval Romance languages. As V2 is asymmetrically distributed in main rather than subordinate clauses, an asymmetry would be expected in phenomena assumed to relate to V2, such as subject inversion, null subject and enclisis. The loss of that asymmetry should therefore indicate the loss of the V2 word order rule. These assumptions are tested here by a quantitative analysis of a treebank of calibrated data covering the crucial period of change (from the 14th to the 16th century) for Medieval French and Venetian. The hard quantitative evidence provided demonstrates that the main versus embedded asymmetry is indeed a micro-cue of V2 structure, and of its loss in one of the two investigated languages.
Quantization is one of the efficient model compression methods, which represents the network with fixed-point or low-bit numbers. Existing quantization methods address the network quantization by treating it as a single-objective optimization that pursues high accuracy (performance optimization) while keeping the quantization constraint. However, owing to the non-differentiability of the quantization operation, it is challenging to integrate the quantization operation into the network training and achieve optimal parameters. In this paper, a novel multi-objective convex quantization for efficient model compression is proposed. Specifically, the network training is modeled as a multi-objective optimization to find the network with both high precision and low quantization error (actually, these two goals are somewhat contradictory and affect each other). To achieve effective multi-objective optimization, this paper designs a quantization error function that is differentiable and ensures the computation convexity in each period, so as to avoid the non-differentiable back-propagation of the quantization operation. Then, we perform a time-series self-distillation training scheme on the multi-objective optimization framework, which distills its past softened labels and combines the hard targets to guarantee controllable and stable performance convergence during training. At last and more importantly, a new dynamic Lagrangian coefficient adaption is designed to adjust the gradient magnitude of quantization loss and performance loss and balance the two losses during training processing. The proposed method is evaluated on well-known benchmarks: MNIST, CIFAR-10/100, ImageNet, Penn Treebank and Microsoft COCO, and experimental results show that the proposed method achieves outstanding performance compared to existing methods.
Natural language processing for Greek and Latin, inflectional languages with small corpora, requires special techniques.For morphological tagging, transformer models show promising potential, but the best approach to use these models is unclear.For both languages, this paper examines the impact of using morphological lexica, training different model types (a single model with a combined feature tag, multiple models for separate features, and a multi-task model for all features), and adding linguistic constraints.We find that, although simply fine-tuning transformers to predict a monolithic tag may already yield decent results, each of these adaptations can further improve tagging accuracy.1 For example, for each type (unique word form) in the GUM English Universal Dependencies Treebank (see https://universaldependencies.org/) there are 10.7 tokens.For the Latin PROIEL treebank there are only 6.5, and for the Greek Perseus treebank even less, viz.4.8 (note that they are all roughly similar in size: 212K, 205K and 202K tokens respectively).
Background: Despite the frequent comorbidity of affective and addictive disorders, the significance of affective dysregulation in problematic pornography use (PPU) is commonly disregarded. The objective of this study is to investigate whether individuals with PPU demonstrate increased sensitivity to negative emotional stimuli in comparison to healthy controls (HCs). Methods: Electrophysiological responses were captured via event-related potentials (ERPs) from 27 individuals with PPU and 29 HCs. They completed an oddball task involving the presentation of deviant stimuli in the form of highly negative (HN), moderately negative (MN), and neutral images, with a standard stimulus being a neutral kettle image. To evaluate participants' subjective feelings of valence and arousal, the Self-Assessment Manikin (SAM) was employed. Results: Regarding subjective evaluations, individuals with PPU indicated diminished valence ratings for HN images as opposed to HCs. Concerning electrophysiological assessments, those with PPU manifested elevated N2 amplitudes in response to both HN and MN images when contrasted against neutral images. Additionally, PPU participants displayed an intensified P3 response to HN images in contrast to MN images, a distinction not evident within the HCs. Discussion: These outcomes suggest that individuals with PPU exhibited heightened reactivity toward negative stimuli. This increased sensitivity to negative cues could potentially play a role in the propensity of PPU individuals to resort to pornography as a coping mechanism for managing stress regulation.
Footwear and its various models are primarily the subject of study in merchandising; however, for linguistics, it is an equally interesting object of research in synchronous and diachronic aspects. In particular, the grammatical characteristics of footwear nouns have not yet been the subject of separate research. The plural form of the noun ʻfootwearʼ is more commonly used in modern linguistic practice, determined by the significance, for human existence, of the semantics of a com- plete pair of footwear. The singular form of the noun name ʻfootwearʼ is less common, as it disrupts the integrity of the concept of a ʻfootwear pairʼ with the semantics of an ʻindi vi- dualizedʼ object. The choice of the number form of footwear nouns as a register is deter- mined by different principles of compiling lexicographical publications and their distinct purposes: explanatory dictionaries prioritize the meaning of an unanalyzed plural, while re- gis ters prioritize the plural form of the noun expressing the semantics of a type of footwear. In orthographic dictionaries, the indication of another numerical form represents the inflec- tional component of the number category of fixed nouns in this semantic group. The grammatical significance of the number in the core of footwear names consists of nouns with a numeral pair, where the codification of the singular form is conditioned by the proportional origin of lexemes and the completion of the grammatical assimilation of foreigh lexemes. Lexicographical sources inconsistently record the singular form of footwear nouns from the core sphere. The change in the grammatical qualification of some footwear names from ʻpluralia tantumʼ to ʻgrammatical significance of singular and pluralʼ attests to the dynamics of morphological norms in contemporary Ukrainian language in general and the process of grammatical adaptation of predominantly foreigh lexemes in particular. However, more often, discrepancies in the printed and ele ctronic dictionaries regarding the grammatical number of nouns from the core sphere of footwear (either plural nouns or nouns with a numeral pair) are caused by subjective factors. The periphery consists of ʻpluralia tantumʼ nouns, most of which are of foreign origin and only make up a part of the lexical composition of the Ukrainian language. The infre- quent functioning of singular forms of nouns from the peripheral sphere in modern speech indicates the incompleteness of their grammatical adaptation. Predominantly foreigh foot- wear names that are not represented in any of the analyzed dictionaries reside outside the delineated spheres, and their use in the plural form in the texts of the 16th version of the Ukrainian National Corpus (HRAK) and in advertising posts on the social network ʻInstagramʼ illustrate the beginning of grammatical adaptation to the morphological norms of the modern Ukrainian language. We see the perspective of research in studying the dynamics of grammatical assimila- tion primarily of those footwear names that are only included in the lexical system of the Ukrainian language and have the grammatical status of ʻpluralia tantumʼ nouns, predomi- nantly used in modern speech either exclusively or mostly in the plural form, with the sin- gular form of these nouns not recorded in any lexicographical publication. The observa- tion of changes in the grammatical number of those footwear names whose singular form is not codified, as it is inconsistently represented in dictionaries, also remains relevant. Keywords: number category, noun, footwear name, pluralia tantum, nu me ral pair, parsed plu- ral, unparsed plural, dynamics, dictionary.
The expression of an association between a conditioned stimulus (CS) and an unconditioned stimulus (US) can be attenuated by presenting the CS by itself (i.e., extinction, Ext). Though effective, Ext is susceptible to recovery effects such as renewal, spontaneous recovery, and reinstatement. Dunsmoor et al. (2015, 2019) have proposed that pairing the CS with a neutral outcome (novelty-facilitated Ext [NFE]) could offer better protection against recovery effects than Ext. Though NFE has been compared to Ext, it has rarely been compared to counterconditioning (CC), a similar procedure except that the CS is paired with a US having a valence opposite to the US used in initial training. We report two aversive conditioning experiments using the rapid-trial streaming procedure with human participants that compare the efficacies and susceptibilities to ABA renewal of Ext, CC, and NFE. Associative learning was assessed through expectancy learning and evaluative conditioning. CC and NFE equally decreased anticipation of the US in the presence of the CS (i.e., expectancy learning). Depending on how the CS-US association was probed, they were either as or more effective at doing so than Ext. All three interference treatments were equally susceptible to context manipulations. Only CC clearly altered the valence of the CS (i.e., evaluative conditioning). Valence ratings after Ext, CC, and NFE, as well as a no-interference control condition, were all equally susceptible to context effects. Overall, the present study does not support the assertion that NFE is consistently more resistant to recovery effects than Ext. (PsycInfo Database Record (c) 2024 APA, all rights reserved).
FrameNet serves as a comprehensive lexical database intended to represent contemporary language usage. However, it faces challenges in accurately representing specialized domains. Among these domains, FrameNet presents difficulties in capturing the specific semantics of human senses. Senses such as smell and taste are in fact included in more general frames or inadequately represented. Building on a previous resource proposing a new framework for olfactory events, we propose a similar annotation scheme for gustatory references in English, enlightening the potential of frames to effectively capture sensory semantics. Having a comprehensive framework to deal with the annotation of this kind of references in textual data is especially important to develop systems for the automatic extraction of sensory information. Moreover, our approach incorporates words from specific historical periods, thereby enriching the framework’s utility for studying language in a diachronic perspective. In this...
The paper presents an overview of initial design choices discussed towards the creation of a treebank for the Italian KIParla corpus
Conventional continuous emotion prediction systems are typically trained to predict the ‘average’ of affect ratings obtained from multiple human annotators. These systems, however, ignore the ambiguity inherent in the perceived emotions, which is not captured by the ‘average rating’. This paper presents a novel ambiguity-aware continuous emotion prediction system that predicts the time-varying emotion state as a series of beta distributions. Our recent work has shown beta distributions to be an effective parametric model of a collection of affect ratings. This work develops an appropriate cost function that enables neural networks to be trained to predict beta distributions. It also investigates the choice of parameterization of the beta distribution, the choice of activation functions of the output layer, and the tractability of gradient definitions in combination with the loss function. The proposed framework is implemented using a Bag-of-Audio-Words front-end and an LSTM-based back-end and evaluated on the RECOLA dataset. In addition to comparison with baseline systems that only predict the ‘average rating’, the effectiveness with which the predictions represent ambiguity in perceived emotions is also evaluated. Experimental results reveal that the proposed approach outperforms other ambiguity-aware systems, especially when predicting valence.
This chapter describes the courageous pedagogical choices made by the author to design and implement a teaching module entitled, “African American English Ain't Broken: The Linguistic Dexterity of Black Folks” for pre-teacher education students enrolled in an multicultural education class in her university's School of Education. The author describes her intention to focus on Black English as a unique variety of English as an act of resistance to white, hegemonic linguistic norms in teacher education. Grounded in Critical Theory and designed using the principles of Culturally Responsive Teaching, this teaching module enables pre-education students to apply their new knowledge of the nature of Black English to develop an understanding of the systems of power and oppression at work in schools and society that marginalize speakers of African American English, including K-12 students. The chapter concludes with suggestions for ways teacher educators, and K-12 classroom teachers can check their linguistic biases and honor students' home languages in meaningful ways.
Abstract Disabled people encounter numerous barriers to accessibility and face discrimination and inequalities in their daily lives. The situation is even more complex for migrants with a disability, who have to learn how to navigate a new bureaucratic system. This study focuses on deaf adult migrants and the linguistic and bureaucratic challenges they face in Swedish society. The data consists of interviews with 43 deaf migrants participating in language learning courses in four folk high schools catering to deaf people in Sweden. Crip Theory and Crip Linguistics are used as lenses to explore the impact of able-bodiedness and linguistic norms on this particular group. The findings show that deaf migrants experience infantilisation, that sign language interpreters are often seen as a one-size-fits-all solution without much consideration for other factors influencing communication, and that normative able-bodiedness underlies many of the bureaucratic issues deaf migrants face.
Pavlovian fear conditioning and extinction represent learning mechanisms underlying exposure-based interventions. While increasing evidence indicates a pivotal role of disgust in the development of contamination-based obsessive-compulsive disorder (C-OCD), dysregulations in conditioned disgust acquisition and maintenance, in particular driven by higher-order conceptual processes, have not been examined. Here, we address this gap by exposing individuals with high (HCC, n = 41) or low (LCC, n = 41) contamination concern to a conceptual-level disgust conditioning and extinction paradigm. Conditioned stimuli (CS+) were images from one conceptual category partially reinforced by unconditioned disgust-eliciting stimuli (US), while images from another category served as non-reinforced conditioned stimuli (CS-). Skin conductance responses (SCRs), US expectancy and CS valence ratings served as primary outcomes to quantify conditioned disgust responses. Relative to LCC, HCC individuals exhibited increased US expectancy and CS+ disgust experience, but comparable SCR levels following disgust acquisition. Despite a decrease in conditioned responses from the acquisition phase to the extinction phase, both groups did not fully extinguish the learned disgust. Importantly, the extinction resilience of acquired disgust was more pronounced in HCC individuals. Together, our findings suggest that individuals with high self-reported contamination concern exhibit increased disgust acquisition and resistance to extinction. The findings provide preliminary evidence on how dysregulated disgust learning mechanism across semantically related concepts may contribute to C-OCD.
Background/Objective: The multidimensional model of the subjective orgasm experience has been validated only in \nthe sexual relationship context, with no evidence for its validity in the solitary masturbation context. This study aims \nto provide validity evidence for this model in the solitary masturbation context by examining the association of its \ndimensions (affective, sensory, intimacy, and rewards) with different sexual arousal measures. Method: Thirty men \nand thirty women viewed content-neutral and sexually explicit masturbation films. Subjective orgasm experience, \npropensity for sexual excitation/inhibition, rating of sexual arousal, rating of genital sensations and genital response \n(penile erection or vaginal pulse amplitude) were assessed. Regression models were conducted to explain the subjective \norgasm experience from sexual arousal measures. Results: Propensity for sexual excitation, propensity for sexual \ninhibition, and the rating of sexual arousal was associated with the different dimensions of the orgasm experience in \nmen, while in women, the rating of sexual arousal and the rating of genital sensations was associated with the sensory \ndimension. Conclusions: Validity evidence is provided for the multidimensional model of the subjective orgasm \nexperience in the solitary masturbation context.
Prosocial and moral behaviors have overlapping neural systems and can both be affected in a number of psychiatric disorders, although whether they involve similar neurochemical systems is unclear. In the current registered randomized placebo-controlled trial on 180 adult male and female subjects, we investigated the effects of intranasal administration of oxytocin and vasopressin, which play key roles in influencing social behavior, on moral emotion ratings for situations involving harming others and on judgments of moral dilemmas where others are harmed for a greater good. Oxytocin, but not vasopressin, enhanced feelings of guilt and shame for intentional but not accidental harm and reduced endorsement of intentionally harming others to achieve a greater good. Neither peptide influenced arousal ratings for the scenarios. Effects of oxytocin on guilt and shame were strongest in individuals scoring lower on the personal distress subscale of trait empathy. Overall, findings demonstrate for the first time that oxytocin, but not vasopressin, promotes enhanced feelings of guilt and shame and unwillingness to harm others irrespective of the consequences. This may reflect associations between oxytocin and empathy and vasopressin with aggression and suggests that oxytocin may have greater therapeutic potential for disorders with atypical social and moral behavior.
Old Permic, also known as Old Komi, is an extinct variety of Komi that was spoken in the late Middle Ages in the lower Vychegda river basin in Northeastern European Russia, in an area that currently is not Komi-speaking. This language variety is attested in fragmentary records from the 14th to 17th century written both in the Old Permic alphabet and in Cyrillic. These records are of significant importance for research on the history of the Komi language. Here we introduce our attempt towards a new Universal Dependencies treebank that will eventually contain the existing corpus of Old Permic in a structured and CoNLL-U annotated format. This will be the first time this material is being made openly available in digital format, and our contribution describes the current state of the art and remaining challenges.
Статтю присвячено розкриттю військових псевдонімів як одного із показників міжособистісної комунікації військовослужбовців у період російсько-української війни 2014–2023 рр. Завдання дослідження – схарактеризувати лексичну базу цих одиниць, розкривши передумови її розвитку та порівнявши з лексичною базою позивних противника. Застосована для аналізу теорія становить взаємодію положень про системність лексики та мовної діяльності, про сутність і функції неофіційного імені на війні у зв’язку зі змістом соціолінгвістичних категорій «неофіційне ім’я», «сленг», «соціогрупа». База даних включає 500 позивних українських учасників російсько-української війни, 500 псевд із минулого століття (для встановлення передумов розвитку лексичної бази сучасних неофіційних імен) та 500 неофіційних імен ворога (для порівняння лексичних баз неофіційних імен українських військовослужбовців і противника). Зіставний та біографічний методи аналізу дозволили отримати нові результати про динаміку лексичної бази позивних у середовищі представників професійно-соціальної групи «українські військовослужбовці». Застосований якісний підхід до лексикологічного аналізу проблеми сприяє формуванню теорії інтерактивної соціолінгвістики.
Abstract The paper analyses the correlation of change in word concreteness ratings with semantic change. To perform the analysis, we apply a neural network to diachronic data to obtain concreteness ratings of English words. As input to the model, we use co-occurrence statistics with the most frequent words extracted from the Google Books Ngram diachronic corpus. It is shown that the model, initially trained on data averaged over a long time interval, predicts the concreteness ratings with high accuracy (based on the word co-occurrence data in a particular year). The impact of lexical semantic change on the change in the concreteness rating is analyzed using 69 words borrowed from previous works. As the considered cases show, the neural network estimate of the word concreteness rating is very sensitive to changes in semantics. Among the factors that influence changes in the concreteness rating, we reveal the emergence of new meanings of a word, the competition of word meanings related to different parts of speech, the use of a word as a proper name, and the use of the word as a part of collocations. It is shown in the paper that changes in the concreteness rating can (along with changes in other word properties) serve as a marker of semantic change.
Emotion recognition from visual stimuli has emerged as a crucial area of research with wide applications in the field of Human-Computer Interaction (HCI) and mental health monitoring. Understanding and predicting emotional responses to visual stimuli from images is a critical task in affective computing. Our study uses deep learning and classical machine learning techniques to classify emotions based on color images. The OASIS image dataset was used; it contains multiple themes of images, including objects, scenes, persons, and animals, with their respective arousal and valence ratings. We applied k-means clustering to identify the number of data points in the maximum cluster within those ratings. We used a Convolutional Neural Network (CNN) regressor for feature extraction of images with their ratings and separately evaluated the error metrics of both the CNN and Random Forest regressor. The results imply that the CNN regression model performs better when predicting emotional dimensions than the Random Forest regression model. This model achieves lower MAE, MSE, and RMSE across the metrics. It shows a more precise and reliable performance in capturing the complexity of emotional dimensions.
How are concepts related to fundamental human experiences organized within the human mind? Our insights are drawn from a semantic network created using the Cross-Linguistic Database of Polysemous Basic Vocabulary, which focuses on a broad range of senses extracted from dictionary entries. The database covers 60 basic vocabularies in 61 languages, providing 11,841 senses from 3736 entries, revealing cross-linguistic semantic connections through automatically generated weighted semantic maps. The network comprises 2941 nodes connected by 3573 edges. The nodes representing body parts, motions, and features closely related to human experience occupy wide fields or serve as crucial bridges across semantic domains in the network. The polysemous network of basic vocabularies across languages represents a shared cognitive network of fundamental human experiences, as these semantic connections should be conceived as generally independent of any specific language and are driven by universal characteristics of the real world as perceived by the human mind. The database holds the potential to contribute to research aimed at unraveling the nature of cognitive proximity.
Natural Language Processing (NLP) has transformed human-machine communication in the digital age, enhancing productivity and unlocking a wealth of possibilities. The effectiveness of NLP hinges on the availability of robust digital resources, such as extensive lexical databases and real-world language corpora. These resources are crucial for various NLP applications, including machine translation, text mining, and speech recognition.NLP's advancements hold immense promise to bridge communication gaps across cultures, provide deeper linguistic insights, and boost productivity across sectors, impacting education, industry, and economic development. However, challenges such as ethical concerns, the necessity for high-quality data, and potential biases in digital language resources must be addressed.This paper presents a vision for the digital resource industry as the cornerstone of NLP, focusing on quantitative transformations that tackle NLP challenges and facilitate big data management. Embracing these transformations, along with a robust digital resource industry, can significantly enhance human-machine interactions and drive future innovations.
Abstract This paper presents the browser-based treebank infrastructure of GLAUx (the Greek Language AUtomated). This linguistic annotation project now has its integrated and user-friendly platform for exploring this data. After discussing the size and types of texts included in the GLAUx corpus, the contribution succinctly surveys the types of linguistic annotation covered by the project (morphology, lemmatization, and syntax). The emphasis of the contribution is on a description of the underlying SQL database structure and the search architecture. Infrastructure-related challenges faced by the GLAUx project are also discussed. Finally, the paper concludes with a discussion of future steps for the project, including additional functionality and expansion of the corpus.
Purpose We explored whether (1) an informational intervention improves ratings of individuals on the autism spectrum (IotAS) in a job interview by curbing salience bias and whether expert-based influence amplifies this effect (Study 1); (2) the effect of disclosure of autism on ratings depends on a candidate’s presentation as IotAS or neurotypical (Studies 1 and 2) and (3) social desirability bias affects ratings of and emotional responses to disclosers (Study 2). Design/methodology/approach In two studies, participants, randomly assigned to experimental conditions, watched a mock job interview of a candidate presenting as an IotAS or neurotypical and reported their perception of his job suitability and selection decision. Study 2 additionally measured participants’ traits associated with social desirability bias, self-reported emotions and involuntary emotions gauged via face-reading software. Findings In Study 1, the informational intervention improved ratings of the IotAS-presenting candidate; delivery by an expert made no difference. Disclosure increased ratings of both the IotAS-presenting and neurotypical-presenting candidates, especially the former, and information mattered more in the absence of disclosure. In Study 2, disclosure improved ratings of the IotAS-presenting candidate only; no evidence of social desirability bias emerged. Originality/value We explain that an informational intervention works by attenuating salience bias, focusing raters on IotAS' qualifications rather than on their unexpected behavior. We also show that disclosure is less helpful for IotAS who behave more neuronormatively and social desirability bias affects neither ratings of nor emotional responses to IotAS-presenting job candidates.
<p>Listening to music often leads to physiological responses. Do these physiological responses contain sufficient information to infer emotion induced in the listener? The current study explores this question by attempting to predict judgments of “felt” emotion from physiological responses alone using linear and neural network models. We measured five channels of peripheral physiology from 20 participants—heart rate (HR), respiration, galvanic skin response, and activity in corrugator supercilii and zygomaticus major facial muscles. Using valence and arousal (VA) dimensions, participants rated their felt emotion after listening to each of 12 classical music excerpts. After extracting features from the five channels, we examined their correlation with VA ratings, and then performed multiple linear regression to see if a linear relationship between the physiological responses could account for the ratings. Although linear models predicted a significant amount of variance in arousal ratings, they were unable to do so with valence ratings. We then used a neural network to provide a non-linear account of the ratings. The network was trained on the mean ratings of eight of the 12 excerpts and tested on the remainder. Performance of the neural network confirms that physiological responses alone can be used to predict musically induced emotion. The non-linear model derived from the neural network was more accurate than linear models derived from multiple linear regression, particularly along the valence dimension. A secondary analysis allowed us to quantify the relative contributions of inputs to the non-linear model. The study represents a novel approach to understanding the complex relationship between physiological responses and musically induced emotion.</p>
This article analyzes the concept of a thesaurus and its capabilities, comparing the differences between the notions of WordNet and a thesaurus. It highlights the structure, composition, and challenges faced in creating linguistic resources like the Turkish WordNet, Arabic WordNet, and the Uzbek UzWordNet, which is based on WordNet. Preliminary linguistic models serve as a foundation for the Uzbek WordNet. Additionally, ideas regarding linguistic research aimed at creating the Uzbek UzNet network are presented.
Emotions arise from a complex interplay of various factors, including conscious experience, physiological processes, and contextual elements. Although emotions are inherently dynamic processes, this aspect is oftentimes neglected in experimental protocols. In this study, we employed dynamical systems theory to investigate the time-varying self-assessed emotion ratings. We used the continuous ratings of the publicly available CASE dataset, in which thirty individuals rated their level of arousal and valence while watching videos designed to evoke four different emotions. Firstly, we analyzed the univariate dynamics by reconstructing the phase space from the arousal and valence series separately, and quantified their regularity and spatial complexity by using three metrics: Fuzzy, Sample, and Distribution Entropy. Then, we combined the arousal and valence series and proposed a novel index, the Multichannel Distribution Entropy (MDistEn), to estimate the complexity of the bivariate phase space. By coupling the two dimensions, we found that MDistEn resulted as an effective marker of fear, showing patterns statistically different from all of the other stimuli (p-value <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\leq$</tex-math></inline-formula> 0.001). These findings support the investigation of the time-varying dynamics of annotated emotion ratings as a promising pathway to discriminate the onset of fear-related pathological states.
Symptom provocation paradigms are paramount to understand a heterogeneous disorder as obsessive-compulsive disorder (OCD). The main aim of our work was to develop and validate an open-access set of OCD-related images comprising three main subtypes: washing, checking, and symmetry. Twenty-six OCD patients and 25 controls provided valence and arousal ratings for a set of OCD-related, aversive, and neutral images. Linear mixed model analyses were used to estimate the main effects of group, image category, and group-image category interaction in image ratings. All main effects were found to be significant for both arousal and valence ratings, except for the group in arousal ratings. Path analysis confirmed our hypothesis that the OCI-R subscales influenced the subjective ratings of the corresponding image categories, particularly among patients. Independent samples t-tests were performed for each OCD picture to compose the set. Arousal demonstrated a greater capacity to distinguish controls and patients, thus sustaining our choice of using these ratings for the final Braga Obsessive-Compulsive Image Set (BOCIS). Our study demonstrated that the stimuli of the BOCIS reliably portray OCD-like triggers for washing, checking and symmetry subtypes. Its open-access availability will facilitate significant progress in both clinical and research settings.
The notion of sound symbolism receives increasing interest in psycholinguistics. Recent research - including empirical effects of affective phonological iconicity on language processing (Adelman et al., 2018; Conrad et al., 2022) - suggested language codes affective meaning at a basic phonological level using specific phonemes as sublexical markers of emotion. Here, in a series of 8 rating-experiments, we investigate the sensitivity of language users to assumed affectively-iconic systematic distribution patterns of phonemes across the German vocabulary:After computing sublexical-affective-values (SAV) concerning valence and arousal for the entire German phoneme inventory according to occurrences of syllabic onsets, nuclei and codas in a large-scale affective normative lexical database, we constructed pseudoword material differing in SAV to test for subjective affective impressions.Results support affective iconicity as affective ratings mirrored sound-to-meaning correspondences in the lexical database. Varying SAV of otherwise semantically meaningless pseudowords altered affective impressions: Higher arousal was consistently assigned to pseudowords made of syllabic constituents more often used in high-arousal words - contrasted by less straightforward effects of valence SAV. Further disentangling specific differential effects of the two highly-related affective dimensions valence and arousal, our data clearly suggest arousal, rather than valence, as the relevant dimension driving affective iconicity effects.
There is great potential for adapting Virtual Reality (VR) exergames based on a user's affective state. However, physical activity and VR interfere with physiological sensors, making affect recognition challenging. We conducted a study (n=72) in which users experienced four emotion inducing VR exergaming environments (happiness, sadness, stress and calmness) at three different levels of exertion (low, medium, high). We collected physiological measures through pupillometry, electrodermal activity, heart rate, and facial tracking, as well as subjective affect ratings. Our validated virtual environments, data, and analyses are openly available. We found that the level of exertion influences the way affect can be recognised, as well as affect itself. Furthermore, our results highlight the importance of data cleaning to account for environmental and interpersonal factors interfering with physiological measures. The results shed light on the relationships between physiological measures and affective states and inform design choices about sensors and data cleaning approaches for affective VR.
Several validated image sets, such as NAPS, IAPS, GAPED, and OASIS, have been developed to elicit affective states. However, these image sets were primarily validated on Western populations within European and American contexts, and none have been fully validated in a Southeast Asian sample, where emotional restraint may also be valued similarly to the East Asian contexts. This study aimed to validate and provide norms for the Nencki Affective Picture System (NAPS; Marchewka et al., 2014) within a Malaysian sample. Subsets from the 1356 NAPS images consisting of five image categories (faces, people, objects, landscapes, animals) were presented sequentially to 409 Malaysian adults aged 18 and above, who rated images for valence, arousal and approach/avoidance on a 9-point Likert scale. Valence, arousal and approach/avoidance norms were compared against the original European sample. Malaysian men and women rated images with lower valence and motivation than Europeans, but Malaysian men showed higher arousal ratings compared to European men, while Malaysian women exhibited the opposite pattern, with lower arousal ratings than European women. A linear regression was found instead of a classic 'boomerang' shaped quadratic regression previously observed in Western samples, suggesting that emotional suppression may be at play, in line with social norms. The Malaysian normative ratings will be freely available to all researchers.
Abstract The present study explores individual differences related to the perception of the media coverage of immigrants as biased. Building on previous research, relations with the extremity in preexisting attitudes toward immigrants and in affective ratings of non-immigrants versus immigrants are examined. Additionally, the present study extends previous work by investigating dogmatism and intellectual humility, including their potential moderating roles on the relations of extremity in attitudes and affective ratings with perceived media bias. A sample of N = 212 (59% men) individuals from the general German population completed self-reports on their preexisting attitudes and affective ratings, as well as dogmatism and intellectual humility online. Moreover, participants rated their perception of news media coverage of immigrants as biased against their views. Results indicate that particularly more extreme negative attitudes toward immigrants and affective ratings favoring non-immigrants relative to immigrants are positively associated with perceiving the media coverage of immigrants as biased. No robust relations of dogmatism, intellectual humility, or their interactions with the extremity scores with perceived media bias were found. These findings underline the importance of negative extremity in (out) group-related attitudes and affect in perceiving the media coverage as biased. New approaches investigating media perceptions from a group-related perspective are discussed.
Phrygian-KUL is a treebank of the ancient Phrygian language for Universal Dependencies (UD). Having originally only annotated the New Phrygian subcorpus, this dataset is continuously being updated to include the entire epigraphic corpus. For more information, please visit the relevant page at the UD project site or the repository on Github.
Cite the source of the dataset as: Kolipakam, Vishnupriya, Michael Dunn, Fiona M. Jordan & Annemarie Verkerk. (2018). DravLex: A Dravidian lexical database. Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands.
Abstract Pictures with affective content have been extensively used in scientific studies of emotion and sexuality. However, only a few standardized picture sets have been developed that offer explicit images, with most lacking pornographic pictures depicting diverse sexual practices. This study aimed to fill this gap through developing a standardized affective set of diverse pornographic pictures (masturbation, oral sex, vaginal sex, anal sex, group sex, paraphilia) of same-sex and opposite-sex content, offering dimensional affective ratings of valence, arousal, and dominance, as well as co-elicited discrete emotions (disgust, moral and ethical acceptance). In total, 192 pornographic pictures acquired from online pornography platforms and 24 control IAPS images have been rated by 319 participants ( M age = 22.66, SD age = 4.66) with self-reported same- and opposite-sex sexual attraction. Stimuli were representative of the entire affective space, including positively and negatively perceived pictures. Participants showed differential affective perception of pornographic pictures according to gender and sexual attraction. Differences in affective ratings related to participants’ gender and sexual attraction, as well as stimuli content (depicted sexual practices and sexes). From the stimuli set, researchers can select explicit pornographic pictures based on the obtained affective ratings and technical parameters (i.e., pixel size, luminosity, color space, contrast, chromatic complexity, spatial frequency, entropy). The stimuli set may be considered a valid tool of diverse explicit pornographic pictures covering the affective space, in particular, for women and men with same- and opposite-sex sexual attraction. This new explicit pornographic picture set (EPPS) is available to the scientific community for non-commercial use.
Emotional experiences deeply impact our bodily states, such as when we feel 'anger', our fists close and our face burns. Recent studies have shown that emotions can be mapped onto specific body areas, suggesting a possible role of the primary somatosensory system (S1) in emotion processing. To date, however, the causal role of S1 in emotion generation remains unclear. To address this question, we applied transcranial alternating current stimulation (tACS) on the S1 at different frequencies (beta, theta, and sham) while participants saw emotional stimuli with different degrees of pleasantness and levels of arousal. Results showed that modulation of S1 influenced subjective emotional ratings as a function of the frequency applied. While theta and beta-tACS made participants rate the emotional images as more pleasant (higher valence), only theta-tACS lowered the subjective arousal ratings (more calming). Skin conductance responses recorded throughout the experiment confirmed a different arousal for pleasant versus unpleasant stimuli. Our study revealed that S1 has a causal role in the feeling of emotions, adding new insight into the embodied nature of emotions. Importantly, we provided causal evidence that beta and theta frequencies contribute differently to the modulation of two dimensions of emotions-arousal and valence-corroborating the view of a dissociation between these two dimensions of emotions.
While artificial Intelligence (AI) has made significant advancements, the seeming absence of its emotional ability has hindered effective communication with humans. This study explores how ChatGPT (ChatGPT-3.5 Mar 23, 2023 Version) represents affective responses to emotional narratives and compare these responses to human responses. Thirty-four participants read affect-eliciting short stories and rated their emotional responses and 10 recorded ChatGPT sessions generated responses to the stories. Classification analyses revealed the successful identification of affective categories of stories, valence, and arousal within and across sessions for ChatGPT. Classification analyses revealed the successful identification of affective categories of stories, valence, and arousal within and across sessions for ChatGPT. Classification accuracies predicting affective categories of stories, valence, and arousal of humans based on the affective ratings of ChatGPT and vice versa were not significant, indicating differences in the way the affective states were represented., indicating differences in the way the affective states were represented. These findings suggested that ChatGPT can distinguish emotional states and generate affective responses consistently, but there are differences in how the affective states are represented between ChatGPT and humans. Understanding these mechanisms is crucial for improving emotional interactions with AI.
Phrygian-KUL is a treebank of the ancient Phrygian language for Universal Dependencies (UD). Having originally only annotated the New Phrygian subcorpus, this dataset is continuously being updated to include the entire epigraphic corpus. For more information, please visit the relevant page at the UD project site or the repository on Github.
The paper presents the tree editor TrEd and related tools that can be used to create, modify, browse, and search treebanks - large language corpora annotated with syntactic and/or semantic structure information. This might include not only phrase structure or dependencies, but also coreference, discourse analysis, and even inter-sentence relations. The project started in the year 2000, and it has been in continuous use since then at various institutions all over the world. Most of the tools are written in Perl, which makes them available to all major operating systems. For searching the treebanks, a query language was developed that describes sets of tree nodes and the relations between them. It also supports aggregation to produce quantitative outputs. There are two different implementations, one translates the queries into SQL statements, the other searches the data directly in the editor. Originally, TrEd supported the PML data format used for the Prague Dependency Treebank. To process data in a different format, one first needed to convert the data into the PML format (and possibly convert the modified data data back to the initial format). Later, a versatile extension system was added to TrEd which made it possible to support other data formats directly. We will show how this works on the example of Universal Dependencies. UD is a framework for grammar annotation across different human languages. The described extension allows TrEd (and some other tools) to open the files in the original UD format natively, building the internal representation on the fly, and also serialise them back after editing.
We introduce Alma ( ﺍ ﺍ ﻠ ى ), an open-source and state-of-the-art lemmatizer, POS tagger, and root tagger for Arabic, boasting both high speed and accuracy. Alma relies on a dictionary of morphological solutions ordered by the frequency of these solutions. This dictionary was developed based on the Qabas lexicographic database. Unlike many Arabic lemmatizers that return a lemma after stripping diacritics, shadda, and hamza (i.e., ambiguous lemma), Alma retrieves unambiguous lemmas (we called true lemmatization). Our POS tagger uses a rich tagset of 40 POS tags. Additionally, our root tagger is the first fully-featured tagger since it uses Qabas, the largest Arabic lexicographic database. We evaluated Alma on the LDC Arabic Treebank (ATB) that contains 339,710 tokens and achieved an 88% F1 score. We also evaluated Alma on the Salma corpus (34k tokens) and obtained a 90% F1 score. Compared to Farasa, MADAMIRA, and Camelira lemmatizers and POS taggers, Alma outperformed all of them in both tasks, excelling in both speed and accuracy. Alma demonstrated superior processing speed, handling 339k tokens in 10.00. Alma is open-source and publicly available at ( https://sina.birzeit.edu/alma ).
Accepting an inflectional subject or null pronoun in impersonal structures with conjugated verbs in Spanish raises several issues. These issues have their origin in a logical approach to grammar and are manifested in the analysis of various types of impersonal structures. To address these challenges, we present a proposal aiming to determine the causes for the obligatory absence of the subject. For that purpose, a distinction is made between a non-existent subject and a hidden or unknown subject. As will be explained, the impossibility of assuming the existence of a subject is due in some cases to lexical restrictions which are carried by certain verbs and verbal complements and are at work at the level of the norm. More specifically, such restrictions are found in constructions with both proper and improper unipersonal verbs and with existential haber. In other cases, the absence of the subject is motivated by the agentive value of specific verbs and the grammatical structure in third person plural structures with unspecified value. Finally, due to specific lexical and grammatical factors, the language system itself prevents the realisation of the constituent in reflexive impersonal structures and in structures employing the modal obligation periphrasis haber que + infinitive. Son varios los problemas que supone la aceptación de un sujeto flexivo o de un pronombre nulo con esa función en estructuras impersonales del español con verbos conjugados. Esos problemas —derivados de una concepción lógica de la oración gramatical y de la generalidad histórica de la gramática— se reflejan en el análisis de diversas estructuras impersonales. Para salvar esos obstáculos, presentamos una propuesta que pretende establecer las causas que provocan la ausencia obligada del sujeto y distinguimos, con este objetivo, un sujeto inexistente y un sujeto oculto o desconocido. Según se detallará, la imposibilidad de suponer la existencia de un sujeto responde a las restricciones léxicas de ciertos verbos y complementos verbales en algunos casos que se sitúan en el plano de la norma (concretamente, en las construcciones con verbos unipersonales propios e impropios y con el existencial haber). Otras veces, el valor agentivo de determinados verbos y la propia estructura gramatical en las construcciones formuladas en tercera persona del plural con valor inespecífico provocan la ausencia del sujeto en este mismo plano. Por último, el propio sistema impide la realización del constituyente en las impersonales reflejas y en las estructuras en las que se emplea la perífrasis modal de obligación haber que + infinitivo por unas razones léxicas y gramaticales concretas. Les problèmes liés à l'acceptation d'un sujet grammatical ou d'un pronom nul dans les structures impersonnelles de l'espagnol avec des verbes conjugués sont multiples. Ces problèmes, découlant de l'identification erronée de la phrase grammaticale avec la prédication logique, de l'indétermination du langage et de la généralité historique de la grammaire, se manifestent dans l'analyse de diverses structures impersonnelles remettant en question la classification de l'espagnol en tant que langue pro-drop prototypique et démontrant la nécessité de développer une hypothèse alternative. Dans cette optique, nous présentons une proposition visant à établir les causes de l'absence obligatoire du sujet et distinguons, à l'intérieur de celle-ci, un sujet inexistant et un sujet caché. Comme cela sera détaillé, cette absence du sujet répond à un critère normatif dans certains cas. Dans d'autres cas, c'est le système lui-même qui empêche la réalisation du constituant.
Abstract: The purpose of the current study was to develop a database of gaming photo stimuli to be used in future psychological research assessing behavioral, cognitive, and neural correlates related to gaming. Participants (ages 18-42, N = 549; 43.17% male) completed ratings on 119 gaming-related images across 5 different categories: valence, arousal, relevance, urge, and interest. A measure of gaming addiction was also included. Positive associations between gaming addiction scores and image ratings were predicted. Gamers rated images higher than non-gamers across multiple dimensions including valence (p =.0012), arousal (p <.0001), urge (p <.0001), and interest (p <.0001). Gaming addiction scores were positively associated with image ratings for valence, r =.399, arousal, r =.438, relevance, r =.215, urge, r =.550, and interest, r =.523, p <.0001. Finally, average image ratings for the overall sample ranged from 5.65 (SD = 2.04) to 3.63 (SD = 1.91) for relevance and interest, respectively. These findings suggest that databases of video gaming imagery, rated for valence, arousal, relevance, urge, and interest, could possibly be used in studies assessing cognitive processing of video gaming-related stimuli in individuals with problematic gaming behavior.
Abstract Persons with dementia are at a higher risk for circadian disturbances and need more bright light to regulate circadian rhythm due to age-related vision deficiencies and reduced activity in suprachiasmatic nuclei. However, the impact of light exposure on nursing home (NH) residents with dementia has not been well evaluated. This secondary analysis examined the associations of daytime light exposure with affect and neurobehavioral symptoms in this population. Circadian stimulus (CS) was included as a lighting measure. Data of 13-week repeated measures were from a clinical trial with 27 residents with dementia. Participants wore light sensors to measure individual light exposure (lux, CCT, and CS) and Philadelphia Affect Rating Scale and Neuropsychiatric Inventory were used to assess affect and neurobehavioral symptoms every other week. Correlation and multilevel modeling (MLM) analyses were performed. Participants’ average age was 87 and 74% were female. Results showed that higher lux levels were significantly correlated with more contentment (r=0.162, p=0.0399) and less sadness (r=-0.185, p=0.0183), less depression (r=-0.157, p=0.0459), and less anxiety (r=-0.207, p=0.0081). Higher CCT levels were significantly correlated with less delusion (r=-0.182, p=0.0205) and appetite changes (r=-0.159, p=0.0434). Higher CS levels were significantly correlated with aberrant motor behavior (r=-0.171, p=0.0293) and anxiety (r=-0.210, p=0.0072). MLM showed that CS significantly predicted decreased aberrant motor behavior (β=-5.04, p&lt; 0.05), nighttime behavior (β=-4.85, p&lt; 0.05), and anxiety (β=-6.08, p&lt; 0.05). The results revealed the importance of light exposure in this population. The findings can guide environmental design to improve affect and neurobehavioral symptoms for NH residents with dementia.
Abstract The development of a benchmark for part-of-speech (PoS) tagging of spoken dialectal European Spanish is presented, which will serve as the foundation for a future treebank. The benchmark is constructed using transcriptions of the Corpus Oral y Sonoro del Español Rural (COSER;“Audible corpus of spoken rural Spanish”) and follows the Universal Dependencies project guidelines. We describe the methodology used to create a gold standard, which serves to evaluate different state-of-the-art PoS taggers (spaCy, Stanza NLP, and UDPipe), originally trained on written data and to fine-tune and evaluate a model for spoken Spanish. It is shown that the accuracy of these taggers drops from 0.98 $$-$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mo>-</mml:mo> </mml:math> 0.99 to 0.94 $$-$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mo>-</mml:mo> </mml:math> 0.95 when tested on spoken data. Of these three taggers, the spaCy’s trf (transformers) and Stanza NLP models performed the best. Finally, the spaCy trf model is fine-tuned using our gold standard, which resulted in an accuracy of 0.98 for coarse-grained tags (UPOS) and 0.97 for fine-grained tags (FEATS). Our benchmark will enable the development of more accurate PoS taggers for spoken Spanish and facilitate the construction of a treebank for European Spanish varieties.
Background While numerous studies observed notable changes in syntactic complexity among individuals with Alzheimer’s disease (AD), these studies predominantly concentrated on isolated internal structures of language but rarely defined the syntactic complexity variation in AD on a continuum. Given that working memory load exhibits a continuous rather than binary pattern in populations and plays a crucial role in syntactic processing and generation, research on syntactic complexity changes in AD could benefit from expanding to include this perspective.Aims To examine the probabilistic distribution of syntactic complexity in AD under controlled sentence length conditions from the perspective of working memory, and to develop a comprehensive linguistic profile of syntactic complexity variation in AD by analyzing fine-grained syntactic features.Methods & procedures The corpus materials consist of descriptions based on the Cookie-Theft picture component provided by 70 individuals with AD (mean MMSE = 20.97) and 70 cognitively intact elderly individuals (mean MMSE = 29.21) from the DementiaBank clinical language transcript dataset. We employed dependency distance to quantify working memory load and syntactic complexity. Additionally, three finer-grained dependency metrics, namely adjacent dependency distribution (1dd%), mean dependency distance (MDD) and dependency direction distribution, enable a deeper interpretation of the syntactic complexity variations in AD from different perspectives.Results (1) The distribution of dependency distances in AD was similar to that in the healthy control (HC) group, both following the Zipf-Alekseev model, and the variations of the parameters within the model were analogous between the two treebanks; (2) The performance of AD patients in adjacent dependency, MDD and dependency direction differed from that of the HC group across varying sentence lengths; (3) “Simplified” and “ungrammatical” structures were the primary syntactic features in AD.Conclusions These findings confirmed the presence of syntactic impairments in AD. The decline of syntactic complexity can be attributed to the failure of the “End Weight” principle due to working memory deficits in AD patients. The study contributes to the easier identification of AD patients and more effective working memory interventions, thereby improving language production in AD patients.