Search

While the two types of intransitive verbs, i.e. unergative and unaccusative, are hypothesised to be syntactically represented, many have proposed a semantic account where abstract properties related to agentivity and telicity, often conceptualised as binary properties, determine the classification. Here we explore the extent to which graded, embodied features rooted in neurobiological systems contribute to the distinction, representing verb meanings as continuous human ratings over various experiential dimensions. Unlike prior studies that classified verbs based on categorical intuition, we assessed the degree of unaccusativity by acceptability of the prenominal past participle construction, one of the unaccusativity diagnostics. Five models were constructed to explain these data: categorical syntactic/semantic, feature-based event-semantic, experiential, and distributional models. The experiential model best explained the diagnostic test data, suggesting that the unaccusative/unergative distinction may be an emergent phenomenon related to differences in underlying experiential content. The experiential model's advantages, including interpretability and scalability, are also discussed. [ABSTRACT FROM AUTHOR]

While research using behavioural methods has repeatedly shown that generic masculines in German come with a male bias, computational methods only entered this area of research very recently. The present paper shows that some assumptions made by these recent computational studies – treating genericity as an inflectional function and computing a vector for generic usage strongly correlated with the grammatical masculine – are not without issue, and offers the use of semantic instance vectors as a possible solution to these issues. Beyond this methodological improvement, the present paper finds that generic masculines are indeed semantically more similar to specific masculines than to specific feminines – results that are in line with findings by the recent computational studies and the majority of previous behavioural studies. [ABSTRACT FROM AUTHOR]

Valence is a dominant semantic dimension, and it is fundamentally linked to basic approach-avoidance behavior within a broad range of contexts. Previous studies have shown that it is possible to approximate the valence of existing words based on several surface-level and semantic components of the stimuli. Parallelly, recent studies have shown that even completely novel and (apparently) meaningless stimuli, like pseudowords, can be informative of meaning based on the information that they carry at the subword level. Here, we aimed to further extend this evidence by investigating whether humans can reliably assign valence to pseudowords and, additionally, to identify the factors explaining such valence judgments. In Experiment 1, we trained several models to predict valence judgments for existing words from their combined form and meaning information. Then, in Experiment 2 and Experiment 3, we extended the results by predicting participants' valence judgments for pseudowords, using a set of models indexing different (possible) sources of valence and selected the best performing model in a completely data-driven procedure. Results showed that the model including basic surface-level (i.e., letters composing the pseudoword) and orthographic neighbors information performed best, thus tracing back pseudoword valence to these components. These findings support perspectives on the nonarbitrariness of language and provide insights regarding how humans process the valence of novel stimuli. [ABSTRACT FROM AUTHOR]

Word affective ratings are important tools in psycholinguistic research, natural language processing, and many other fields. However, even for well-studied languages, such norms are usually limited in scale. To extrapolate affective (i.e., valence and arousal) values for words in the SUBTLEX-CH database (Cai & Brysbaert, 2010, PLoS ONE, 5(6):e10729), we implemented a computational neural network which captured how words' vector-based semantic representations corresponded to the probability densities of their valence and arousal. Based on these probability density functions, we predicted not only a word's affective values, but also their respective degrees of variability that could characterize individual differences in human affective ratings. The resulting estimates of affective values largely converged with human ratings for both valence and arousal, and the estimated degrees of variability also captured important features of the variability in human ratings. We released the extrapolated affective values, together with their corresponding degrees of variability, for over 38,000 Chinese words in the Open Science Framework (https://osf.io/s9zmd/). We also discussed how the view of embodied cognition could be illuminated by this computational model. [ABSTRACT FROM AUTHOR]

AI language models are increasingly transforming language research in various ways. How can language educators and researchers respond to the challenge posed by these AI models? Specifically, how can we embrace this technology to inform and enhance second language learning and teaching? In order to quantitatively characterize and index second language writing, the current work proposes the use of similarities derived from contextualized meaning representations in AI language models. The computational analysis in this work is hypothesis-driven. The current work predicts how similarities should be distributed in a second language learning setting. The results suggest that similarity metrics are informative of writing proficiency assessment and interlanguage development. Statistically significant effects were found across multiple AI models. Most of the metrics could distinguish language learners' proficiency levels. Significant correlations were also found between similarity metrics and learners' writing test scores provided by human experts in the domain. However, not all such effects were strong or interpretable. Several results could not be consistently explained under the proposed second language learning hypotheses. Overall, the current investigation indicates that with careful configuration and systematic metrics design, AI language models can be promising tools in advancing language education. [ABSTRACT FROM AUTHOR]

Background: This study aims to facilitate the identification of similar devices for both, the European Medical Device Regulation (MDR) and the US 510(k) equivalence pathway by leveraging existing data. Both are related to the regulatory pathway of read across for chemicals, where toxicological data from a known substance is transferred to one under investigation, as they aim to streamline the accreditation process for new devices and chemicals. Research design and methods: This study employs latent semantic analysis to generate similarity values, harnessing the US Food and Drug Administration 510k-database, utilizing their 'Device Descriptions' and 'Intended Use' statements. Results: For the representative inhaler cluster, similarity values up to 0.999 were generated for devices within a 510(k)-predicate tree, whereas values up to 0.124 were gathered for devices outside this group. Conclusion: Traditionally, MDR equivalence involves manual review of many devices, which is laborious. However, our results suggest that the automated calculation of similarity coefficients streamlines this process, thus reducing regulatory effort, which can be beneficial for patients needing medical devices. Although this study is focused on the European perspective, it can find application within 510(k) equivalence regulation. The conceptual approach is reminiscent of chemical fingerprint similarity analysis employed in read-across. Plain Language Summary: This study addresses improvement of the registration process for medical devices by using automated methods to determine how similar they are to existing devices. Such a process is already used in chemistry for analysis of related substances. In the context of Medical Device Regulation (MDR), which sets standards for these devices, this process might be applicable in device equivalence evaluation. Traditionally, proving equivalence involves manually finding devices that are similar, but this is time-consuming, repetitive and labor-intensive. This study proposes a new approach, using advanced computer methods and a database from the US Food and Drug Administration (FDA) to automatically identify similar devices. This could make the process much quicker and more accurate and furthermore reduce bias. The study suggests that by applying these automated methods, the impact of recent regulatory changes could be reduced. This means that proving equivalence, a critical step to facilitate device accreditation, could be done more efficiently. The study shows potential for a significant transformation in compliance processes within the medical device industry, making them more streamlined and automated. [ABSTRACT FROM AUTHOR]

This paper presents a large and systematic assessment of distributional models for Galician. To this end, we have first trained and evaluated static word embeddings (e.g., word2vec, GloVe), and then compared their performance with that of current contextualised representations generated by neural language models. First, we have compiled and processed a large corpus for Galician, and created four datasets for word analogies and concept categorisation based on standard resources for other languages. Using the aforementioned corpus, we have trained 760 static vector space models which vary in their input representations (e.g., adjacency-based versus dependency-based approaches), learning algorithms, size of the surrounding contexts, and in the number of vector dimensions. These models have been evaluated both intrinsically, using the newly created datasets, and on extrinsic tasks, namely on POS-tagging, dependency parsing, and named entity recognition. The results provide new insights into the performance of different vector models in Galician, and about the impact of several training parameters on each task. In general, fastText embeddings are the static representations with the best performance in the intrinsic evaluations and in named entity recognition, while syntax-based embeddings achieve the highest results in POS-tagging and dependency parsing, indicating that there is no significant correlation between the performance in the intrinsic and extrinsic tasks. Finally, we have compared the performance of static vector representations with that of BERT-based word embeddings, whose fine-tuning obtains the best performance on named entity recognition. This comparison provides a comprehensive state-of-the-art of current models in Galician, and releases new transformer-based models for NER. All the resources used in this research are freely available to the community, and the best models have been incorporated into SemantiGal, an online tool to explore vector representations for Galician. [ABSTRACT FROM AUTHOR]

Using distributional semantics, we show that English nominal pluralization exhibits semantic clusters. For instance, the change in semantic space from singulars to plurals differs depending on whether a word denotes, e.g., a fruit, or an animal. Languages with extensive noun classes such as Swahili and Kiowa distinguish between these kind of words in their morphology. In English, even though not marked morphologically, plural semantics actually also varies by semantic class. A semantically informed method, CosClassAvg, is introduced that is compared to two other methods, one implementing a fixed shift from singular to plural, and one creating plural vectors from singular vectors using a linear mapping (FRACSS). Compared to FRACSS, CosClassAvg predicted plural vectors that were more similar to the corpus-extracted plural vectors in terms of vector length, but somewhat less similar in terms of orientation. Both FRACSS and CosClassAvg outperform the method using a fixed shift vector to create plural vectors, which does not do justice to the intricacies of English plural semantics. A computational modeling study revealed that the observed difference between the plural semantics generated by these three methods carries over to how well a computational model of the listener can understand previously unencountered plural forms. Among all methods, CosClassAvg provides a good balance for the trade-off between productivity (being able to understand novel plural forms) and faithfulness to corpus-extracted plural vectors (i.e., understanding the particulars of the meaning of a given plural form). [ABSTRACT FROM AUTHOR]

Do large language models like Chat-GPT or Claude meaningfully use the words they produce? Or are they merely clever prediction machines, simulating language use by producing statistically plausible text? There have already been some initial attempts to answer this question by showing that these models meet the criteria for entering meaningful states according to metasemantic theories of mental content. In this paper, I will argue for a different approach—that we should instead consider whether language models meet the criteria given by our best metasemantic theories of linguistic content. In that vein, I will illustrate how this can be done by applying two such theories to the case of language models: Gareth Evans’ (1982) account of naming practices and Ruth Millikan’s (1984, 2004, 2005) teleosemantics. In doing so, I will argue that it is a mistake to think that the failure of LLMs to meet plausible conditions for mental intentionality thereby renders their outputs meaningless, and that a distinguishing feature of linguistic intentionality—dependency on a pre-existing linguistic system—allows for the plausible result that LLM outputs are meaningful. [ABSTRACT FROM AUTHOR]

We investigate the interface between concepts and word meanings by asking English speakers to list members of superordinate categories under one of three conditions: (1) when cued by a label (e.g. animals), (2) an exemplar list (e.g. dog, cat, mouse), or (3) a definition (e.g. "living creatures that roam the Earth"). We find that categories activated by labels lead to participants listing more category-typical responses, as quantified through typicality ratings, similarity in word embedding space, and accuracy in guessing category labels. This effect is stronger for some categories than others (e.g. stronger for appetizers than animals). These results support the view that a word is not merely a label for a concept, but rather a unique way of accessing and organizing conceptual space. [ABSTRACT FROM AUTHOR]

The formation of false memories is one of the most widely studied topics in cognitive psychology. The Deese–Roediger–McDermott (DRM) paradigm is a powerful tool for investigating false memories and revealing the cognitive mechanisms subserving their formation. In this task, participants first memorize a list of words (encoding phase) and next have to indicate whether words presented in a new list were part of the initially memorized one (recognition phase). By employing DRM lists optimized to investigate semantic effects, previous studies highlighted a crucial role of semantic processes in false memory generation, showing that new words semantically related to the studied ones tend to be more erroneously recognized (compared to new words less semantically related). Despite the strengths of the DRM task, this paradigm faces a major limitation in list construction due to its reliance on human-based association norms, posing both practical and theoretical concerns. To address these issues, we developed the False Memory Generator (FMG), an automated and data-driven tool for generating DRM lists, which exploits similarity relationships between items populating a vector space. Here, we present FMG and demonstrate the validity of the lists generated in successfully replicating well-known semantic effects on false memory production. FMG potentially has broad applications by allowing for testing false memory production in domains that go well beyond the current possibilities, as it can be in principle applied to any vector space encoding properties related to word referents (e.g., lexical, orthographic, phonological, sensory, affective, etc.) or other type of stimuli (e.g., images, sounds, etc.). [ABSTRACT FROM AUTHOR]

Language is inherently multimodal. In spoken languages, combined spoken and visual signals (e.g., co‐speech gestures) are an integral part of linguistic structure and language representation. This requires an extension of the parallel architecture, which needs to include the visual signals concomitant to speech. We present the evidence for the multimodality of language. In addition, we propose that distributional semantics might provide a format for integrating speech and co‐speech gestures in a common semantic representation. [ABSTRACT FROM AUTHOR]

We present a word similarity gold standard for Akkadian, a language documented in ancient Mesopotamian sources from the 24th century BCE until the first century CE. The gold standard comprises 300 word pairs ranked by their paradigmatic similarity by five independently working Assyriologists. We use the gold standard to tune PMI + SVD and fastText models to improve their performance. We also present a hyper-parametrized PMI + SVD model for building count-based word embeddings, that aims to deal with the data sparsity and repetition issues encountered in Akkadian texts. Our model combines Dirichlet smoothing with context distribution smoothing, and uses context similarity weighting to down-sample distortion caused by formulaic litanies and partially or fully duplicated passages. [ABSTRACT FROM AUTHOR]

Word frequency is a strong predictor in most lexical processing tasks. Thus, any model of word recognition needs to account for howword frequency effects arise. The Discriminative Lexicon Model (DLM)models lexical processing withmappings between words' forms and their meanings. Comprehension and production are modeled via linear mappings between the two domains. So far, the mappings within the model can either be obtained incrementally via error-driven learning, a computationally expensive process able to capture frequency effects, or in an efficient, but frequency-agnostic solution modeling the theoretical endstate of learning (EL) where all words are learned optimally. In the present study we show how an efficient, yet frequency-informed mapping between form and meaning can be obtained (Frequency-informed learning; FIL). We find that FIL well approximates an incremental solution while being computationallymuch cheaper. FIL shows a relatively low type-and high token-accuracy, demonstrating that the model is able to process most word tokens encountered by speakers in daily life correctly. We use FIL to model reaction times in the Dutch Lexicon Project by means of a Gaussian Location Scale Model and find that FIL predicts well the S-shaped relationship between frequency and the mean of reaction times but underestimates the variance of reaction times for low frequency words. FIL is also better able to account for priming effects in an auditory lexical decision task in Mandarin Chinese, compared to EL. Finally, we used ordered data from CHILDES to compare mappings obtained with FIL and incremental learning. We show that themappings are highly correlated, but that with FIL some nuances based on word ordering effects are lost. Our results show how frequency effects in a learning model can be simulated efficiently, and raise questions about how to best account for low-frequency words in cognitive models. [ABSTRACT FROM AUTHOR]

We present a computational model capable of simulating aspects of human knowledge for thousands of real-world concepts. Our approach involves a pretrained transformer network that is further fine-tuned on large data sets of participant-generated feature norms. We show that such a model can successfully extrapolate from its training data, and predict human knowledge for new concepts and features. We apply our model to stimuli from 25 previous experiments in semantic cognition research and show that it reproduces many findings on semantic verification, concept typicality, feature distribution, and semantic similarity. We also compare our model against several variants, and by doing so, establish the model properties that are necessary for good prediction. The success of our approach shows how a combination of language data and (laboratory-based) psychological data can be used to build models with rich world knowledge. Such models can be used in the service of new psychological applications, such as the modeling of naturalistic semantic verification and knowledge retrieval, as well as the modeling of real-world categorization, decision-making, and reasoning. [ABSTRACT FROM AUTHOR]

In order to study global experience for legislation changing and rule-making necessitates, tools for information retrieval of regulatory documents written in different languages become increasingly necessary. One of the aspects of information identification is retrieval of thematically similar documents for a given input document. In this context, an important task of cross-lingual search arises when the user of an information system specifies a reference document in one language, and the search results contain relevant documents in other languages. The article describes different approaches to solving this problem: from classic mediator-based methods to more modern solutions, based on distributional semantics. The test collection used in the study was taken from the United Nations Digital Library, which provides legal documents in both the original English and their Russian translations. [ABSTRACT FROM AUTHOR]

Using a large data set of online media content in eight European countries, this paper broadens the empirical investigation of the online hostility hypothesis, which posits that interactions on social sites such as blogs and forums contain more hostile expressions toward minority groups than social interactions offline or in editorial news media. Overall, our results are consistent with the online hostility hypothesis when comparing news media content with social sites, but we find that negatively charged representations are common in both media types. It is instead the amount of attention to Muslims and Islam on social sites that most clearly differs and is the main driver of online hostility in the online media environment more broadly conceived. [ABSTRACT FROM AUTHOR]

AI language models are increasingly transforming language research in various ways. How can language educators and researchers respond to the challenge posed by these AI models? Specifically, how can we embrace this technology to inform and enhance second language learning and teaching? In order to quantitatively characterize and index second language writing, the current work proposes the use of similarities derived from contextualized meaning representations in AI language models. The computational analysis in this work is hypothesis-driven. The current work predicts how similarities should be distributed in a second language learning setting. The results suggest that similarity metrics are informative of writing proficiency assessment and interlanguage development. Statistically significant effects were found across multiple AI models. Most of the metrics could distinguish language learners’ proficiency levels. Significant correlations were also found between similarity metrics and learners’ writing test scores provided by human experts in the domain. However, not all such effects were strong or interpretable. Several results could not be consistently explained under the proposed second language learning hypotheses. Overall, the current investigation indicates that with careful configuration and systematic metrics design, AI language models can be promising tools in advancing language education.

ABSTRACTUsing a large data set of online media content in eight European countries, this paper broadens the empirical investigation of the online hostility hypothesis, which posits that interactions on social sites such as blogs and forums contain more hostile expressions toward minority groups than social interactions offline or in editorial news media. Overall, our results are consistent with the online hostility hypothesis when comparing news media content with social sites, but we find that negatively charged representations are common in both media types. It is instead the amount of attention to Muslims and Islam on social sites that most clearly differs and is the main driver of online hostility in the online media environment more broadly conceived.

Lexical ambiguity is pervasive in language, and the nature of the representations of an ambiguous word's multiple meanings is yet to be fully understood. With a special focus on Chinese characters, the present study first established that native speaker's perception about a character's number of meanings was heavily influenced by the availability of its distinct word formations, while whether these meanings would be perceived to be closely related was driven by further conceptual analysis. These notions were operationalized as two computed metrics, which assessed the degree of dispersion across individual word formations and the degree of propinquity across clusters of word formations, respectively, in a distributional semantic space. The observed correlations between the computed and the perceived metrics indicated that the utility of word formations to tap into meaning representations of Chinese characters was indeed cognitively plausible. The results have demonstrated the extent to which distributional semantics could inform about meaning representations of Chinese characters, which has theoretical implications for the representation of ambiguous words more generally. [ABSTRACT FROM AUTHOR]

Free association among words is a fundamental and ubiquitous memory task. Although distributed semantics (DS) models can predict the association between pairs of words, and semantic network (SN) models can describe transition probabilities in free association data, there have been few attempts to apply established cognitive process models of memory search to free association data. Thus, researchers are currently unable to explain the dynamics of free association using memory mechanisms known to be at play in other retrieval tasks, such as free recall from lists. We address this issue using a popular neural network model of free recall, the context maintenance and retrieval (CMR) model, which we fit using stochastic gradient descent on a large data set of free association norms. Special cases of CMR mimic existing DS and SN models of free association, and we find that CMR outperforms these models on out-of-sample free association data. We also show that training CMR on free association data generates improved predictions for free recall from lists, demonstrating the value of free association for the study of many different types of memory phenomena. Overall, our analysis provides a new account of the dynamics of free association, predicts free association with increased accuracy, integrates theories of free association with established models of memory, and shows how large data sets and neural network training methods can be used to model complex cognitive processes that operate over thousands of representations. [ABSTRACT FROM AUTHOR]

The tradition of dimensional models in the study of emotions suggests that affective space is better defined by a small number of non-specific general dimensions. This dimensional perspective in the study of emotions postulates that the minimal entities of representation are dimensions such as valence (attraction vs. rejection) and arousal (level of activation). Data collection in this perspective is done through the laborious process of resorting to the estimates of hundreds of people who must decide on a continuum of two dimensions: how positive or negative the object to which the concept alludes is, and what level of arousal it generates. To do so, generally, an image-based questionnaire developed to measure an emotional response is used, called SAM (Self Assessment Manikin), which is a set of synthesized drawings that can be used to guide the participants' response. As can be noted, the volume and quality of the procedures used to study these affective variables associated with concepts involves an arduous process of data collection and processing. The complexity of this work, due to the enormous amount and type of data, makes computational intelligence a very useful and novel tool for its approach in order to obtain reliable and reproducible results. Processing by means of distributional semantics obtains the meaning of a word by locating the context in which it appears through an intelligent search in large volumes of electronically stored data. Lexical-affective information, represented by the valence and activation dimensions, is a type of information that seems to be represented by the social use of a term and, therefore, it is plausible to infer it by studying the contexts in which the term appears. In this paper we consider the plausibility of the application of the method of estimating lexical-affective values by means of distributional semantics for the Spanish language. In order to achieve this objective, the Spanish adaptation of the ANEW is used for this purpose and, two procedures were carried out. On the one hand, the most traditionally used computational method of estimation by means of linear regressions was compared with a model based on neural networks, which showed the better fit of the latter. On the other hand, lexical-affective values were estimated for a set of complex emotions and, in order to verify the strength of the results, such estimation was compared with empirical data taken and processed in a linguistic community (Argentine), different from the community that gave rise to the data with which the computational model was trained (Spanish). The results have been very encouraging since, between the computationally derived estimates and the empirically derived data with the Argentine population, correlations were found to be sufficiently strong and comparable to those that can be found when the same comparison is made between empirically derived results with different age groups, or between different genders, or when comparing different geographical regions. These results point to a striking common base of the language, a basic common core of concepts, which can be explored by means of the procedures and techniques of distributional semantics. [ABSTRACT FROM AUTHOR]

This replication study aims to investigate a potential bias toward addition in the German language, building upon previous findings of Winter and colleagues who identified a similar bias in English. Our results confirm a bias in word frequencies and binomial expressions, aligning with these previous findings. However, the analysis of distributional semantics based on word vectors did not yield consistent results for German. Furthermore, our study emphasizes the crucial role of selecting appropriate translational equivalents, highlighting the significance of considering language‐specific factors when testing for such biases for languages other than English. [ABSTRACT FROM AUTHOR]

This article focuses on two research questions arising from the 2010 U.S. Securities and Exchange Commission (SEC) Advisory on climate change reporting: (1) How does the discussion of climate change in SEC filings change after the Advisory? and (2) What are firms talking about when they talk about climate change? Findings were obtained from the 218,000 10‐K filings to the SEC during the 2000–2019 period. The study develops and applies text mining methodology based on extracting information from the "semantic associates" in the "neighborhoods" of indicative terms. On (1) it finds that climate change‐related reporting does increase substantially after the SEC guidance. On (2) a nuanced picture emerges. Firms with comparatively larger transition risks tend to discuss climate change comparatively more, focusing on regulation‐related topics. Firms exposed to the physical risks of climate change tend to discuss climate change somewhat less, focusing on meteorological topics. The results enrich our understanding regarding environmental policies and firms' behaviors regarding climate change. Theoretical and practical implications are provided. [ABSTRACT FROM AUTHOR]

Unsupervised image translation (UIT) studies the mapping between two image domains. Since such mappings are under-constrained, existing research has pursued various desirable properties such as distributional matching or two-way consistency. In this paper, we re-examine UIT from a new perspective: distributional semantics consistency, based on the observation that data variations contain semantics, e.g., shoes varying in colors. Further, the semantics can be multi-dimensional, e.g., shoes also varying in style, functionality, etc. Given two image domains, matching these semantic dimensions during UIT will produce mappings with explicable correspondences, which has not been investigated previously. We propose distributional semantics mapping (DSM), the first UIT method which explicitly matches semantics between two domains. We show that distributional semantics has been rarely considered within and beyond UIT, even though it is a common problem in deep learning. We evaluate DSM on several benchmark datasets, demonstrating its general ability to capture distributional semantics. Extensive comparisons show that DSM not only produces explicable mappings, but also improves image quality in general. [ABSTRACT FROM AUTHOR]

Abstract Unsupervised image translation (UIT) studies the mapping between two image domains. Since such mappings are under-constrained, existing research has pursued various desirable properties such as distributional matching or two-way consistency. In this paper, we re-examine UIT from a new perspective: distributional semantics consistency, based on the observation that data variations contain semantics, e.g., shoes varying in colors. Further, the semantics can be multi-dimensional, e.g., shoes also varying in style, functionality, etc. Given two image domains, matching these semantic dimensions during UIT will produce mappings with explicable correspondences, which has not been investigated previously. We propose distributional semantics mapping (DSM), the first UIT method which explicitly matches semantics between two domains. We show that distributional semantics has been rarely considered within and beyond UIT, even though it is a common problem in deep learning. We evaluate DSM on several benchmark datasets, demonstrating its general ability to capture distributional semantics. Extensive comparisons show that DSM not only produces explicable mappings, but also improves image quality in general.

Most words are low in frequency, yet a prevailing theory of word meaning (the distributional hypothesis: that words with similar meanings occur in similar contexts) and corresponding computational models struggle to represent low-frequency words. We conducted two preregistered experiments to test the hypothesis that similar-sounding words flesh out deficient semantic representations. In Experiment 1, native English speakers made semantic relatedness decisions about a cue (e.g., dodge) followed either by a target that overlaps in form and meaning with a higher frequency word (evade, which overlaps with avoid) or by a control (elude), matched on distributional and formal similarity to the cue. (Participants did not see higher frequency words like avoid.) As predicted, participants decided faster and more often that overlapping targets, compared to controls, were semantically related to cues. In Experiment 2, participants read sentences containing the same cues and targets (e.g., The kids dodged something and She tried to evade/elude the officer). We used MouseView.js to blur the sentences and create a fovea-like aperture directed by the participant's cursor, allowing us to approximate fixation duration. While we did not observe the predicted difference at the target region (e.g., evade/elude), we found a lag effect, with shorter fixations on words following overlapping targets, suggesting easier integration of those meanings. These experiments provide evidence that words with overlapping forms and meanings bolster representations of low-frequency words, which supports approaches to natural language processing that incorporate both formal and distributional information and which revises assumptions about how an optimal language will evolve. [ABSTRACT FROM AUTHOR]

Recent years have seen a flourishing of Natural Language Processing models that can mimic many aspects of human language fluency. These models harness a simple, decades‐old idea: It is possible to learn a lot about word meanings just from exposure to language, because words similar in meaning are used in language in similar ways. The successes of these models raise the intriguing possibility that exposure to word use in language also shapes the word knowledge that children amass during development. However, this possibility is strongly challenged by the fact that models use language input and learning mechanisms that may be unavailable to children. Across three studies, we found that unrealistically complex input and learning mechanisms are unnecessary. Instead, simple regularities of word use in children's language input that they have the capacity to learn can foster knowledge about word meanings. Thus, exposure to language may play a simple but powerful role in children's growing word knowledge. A video abstract of this article can be viewed at https://youtu.be/dT83dmMffnM. Research Highlights: Natural Language Processing (NLP) models can learn that words are similar in meaning from higher‐order statistical regularities of word use.Unlike NLP models, infants and children may primarily learn only simple co‐occurrences between words.We show that infants' and children's language input is rich in simple co‐occurrence that can support learning similarities in meaning between words.We find that simple co‐occurrences can explain infants' and children's knowledge that words are similar in meaning. [ABSTRACT FROM AUTHOR]

So-called "distributional" language models have become dominant in research on the computational modelling of lexical semantics. This paper investigates how well such models perform on Ancient Greek, a highly inflected historical language. It compares several ways of computing such distributional models on the basis of various context features (including both bag-of-words features and syntactic dependencies). The performance is assessed by evaluating how well these models are able to retrieve semantically similar words to a given target word, both on a benchmark we designed ourselves as well as on several independent benchmarks. It finds that dependency features are particularly useful to calculate distributional vectors for Ancient Greek (although the level of granularity that these dependency features should have is still open to discussion) and discusses possible ways for further improvement, including addressing problems related to polysemy and genre differences. [ABSTRACT FROM AUTHOR]

Word frequency (WF) is a strong predictor of lexical behavior. However, much research has shown that measures of contextual and semantic diversity offer a better account of lexical behaviors than WF (Adelman et al., 2006; Jones et al., 2012). In contrast to these previous studies, Chapman and Martin (2022) recently demonstrated that WF seems to account for distinct and greater levels of variance than measures of contextual and semantic diversity across a variety of datatypes. However, there are two limitations to these findings. The first is that Chapman and Martin (2022) compared variables derived from different corpora, which makes any conclusion about the theoretical advantage of one metric over another confounded, as it could be the construction of one corpus that provides the advantage and not the underlying theoretical construct. Second, they did not consider recent developments in the semantic distinctiveness model (SDM; Johns, 2021a; Johns et al., 2020; Johns & Jones, 2022). The current paper addressed the second limitation. Consistent with Chapman and Martin (2022), our results showed that the earliest versions of the SDM were less predictive of lexical data relative to WF when derived from a different corpus. However, the later versions of the SDM accounted for substantially more unique variance than WF in lexical decision and naming data. The results suggest that context-based accounts provide a better explanation of lexical organization than repetition-based accounts. [ABSTRACT FROM AUTHOR]

The purpose of this work is to investigate the problem of detecting transportable borrowings and text reuse. The article proposes a monolingual solution to this problem: translating the suspicious material into language collections for additional monolingual analysis. One of the major requirements for the suggested technique is robustness against machine learning ambiguities. The next step in the document analysis is split into two parts. The authors begin by retrieving documents-candidates that are similarity to other types of text recurrence. The paper proposes retrieving texts utilizing word clusters formed using distributional semantic for robustness. In the second stage, the authors use deep learning neural networks to compare the suspected document to candidates utilizing phrase embedding. The experimentation is carried out for the language pair "English-Arabic" on both articles and synthetic data. [ABSTRACT FROM AUTHOR]

This paper seeks to address the relation between semantics, pragmatics and the productivity of a low level lexico-grammatical construction, Have the N (body part/attitude) to ↔metapragmatic comment. The question posed is how semantics affects productivity, in the generative sense of extensibility of a construction (a form meaning pairing). The method identifies the specificity and variations of the Have the N (body part/attitude) to ↔metapragmatic comment construction within the pragmeme of politeness using the COHA. Hereafter, we consider how to measure the extensibility within the onomasiological frame based on the available pool of forms expressing an attitude/emotion, i.e., the coverage or attractivity of the Have the N to construction. The paper discusses the findings, namely, how to overcome methodological issues relating to a qualitative rather than quantitative approach to the constructional architecture and the relative productivity of constructions. The experimental small scale corpus study of Have the N to in the COHA suggests that a global view of constructional architecture at multiple levels should be pertinent to identifying the extensibility potential of the construction. [ABSTRACT FROM AUTHOR]

This paper provides a comparative analysis of cross-lingual word embedding by studying the impact of different variables on the quality of the embedding models within the distributional semantics framework. Distributional semantics is a method for the semantic representation of words, phrases, sentences, and documents. This method aims at capturing as much information as possible from the contextual information in a vector space. The early study in this domain focused on monolingual word embedding. Further progress used cross-lingual data to capture the contextual semantic information across different languages. The main contribution of this research is to make a comparative study to find out the superior impact of the learning methods, supervised and unsupervised in training and post-training approaches in different embedding algorithms, to capture semantic properties of the words in cross-lingual embedding models to be applicable in tasks that deal with multi-languages, such as question retrieval. To this end, we study the cross-lingual embedding models created by BilBOWA, VecMap, and MUSE embedding algorithms along with the variables that impact the embedding models' quality, namely the size of the training data and the window size of the local context. In our study, we use the unsupervised monolingual Word2Vec embedding model as the baseline and evaluate the quality of embeddings on three data sets: Google analogy, mono- and cross-lingual words similar lists. We further investigated the impact of the embedding models in the question retrieval task.

Linguistic Explorations of Societies (LES) is an interdisciplinary research project with scholars from the fields of political science, computer science, and computational linguistics. The overarching ambition of LES has been to contribute to the survey-based comparative scholarship by compiling and analyzing online text data within and between languages and countries. To this end, the project has developed an online semantic lexicon, which allows researchers to explore meanings and usages of words in online media across a substantial number of geo-coded languages. The lexicon covers data from approximately 140 language–country combinations and is, to our knowledge, the most extensive free research resource of its kind. Such a resource makes it possible to critically examine survey translations and identify discrepancies in order to modify and improve existing survey methodology, and its unique features further enable Internet researchers to study public debate online from a comparative perspective. In this article, we discuss the social scientific rationale for using online text data as a complement to survey data, and present the natural language processing–based methodology behind the lexicon including its underpinning theory and practical modeling. Finally, we engage in a critical reflection about the challenges of using online text data to gauge public opinion and political behavior across the world. [ABSTRACT FROM AUTHOR]

This work aims to study the construction of migrant categories and immigration discourse on Swedish-speaking Facebook pages in the last decade. It combines the insights from computational linguistics and distributional semantics approach with those from discursive psychology to explore a corpus of more than 1 M Facebook posts. This allows one to compare the meanings of labels denoting various categories of migrants and identify the key interpretative repertoires used to discuss the immigration topic. The study finds that the 'immigrant' category has stronger association with potential costs, benefits and threat to the host society, while the 'refugee' category is presented as in need of support and solidarity. Nevertheless, objectification and exclusionary rhetoric are used in relation to both categories, although in different ways, while the immigration issue is often interpreted as a matter of Sweden's national concern rather than as a part of people's actual experiences and life paths. [ABSTRACT FROM AUTHOR]

Machines have achieved a broad and growing set of linguistic competencies, thanks to recent progress in Natural Language Processing (NLP). Psychologists have shown increasing interest in such models, comparing their output to psychological judgments such as similarity, association, priming, and comprehension, raising the question of whether the models could serve as psychological theories. In this article, we compare how humans and machines represent the meaning of words. We argue that contemporary NLP systems are fairly successful models of human word similarity, but they fall short in many other respects. Current models are too strongly linked to the text-based patterns in large corpora, and too weakly linked to the desires, goals, and beliefs that people express through words. Word meanings must also be grounded in perception and action and be capable of flexible combinations in ways that current systems are not. We discuss promising approaches to grounding NLP systems and argue that they will be more successful, with a more human-like, conceptual basis for word meaning. [ABSTRACT FROM AUTHOR]

We contrast two views of rivalry in word formation. Under the classical, categorical view, two processes are rivals if they are semantically equivalent. Under the more nuanced, gradient view, two processes can be rivals at different degrees, depending on how frequently they are amenable to be deployed as alternatives to one another. We propose to use methods from distributional semantics to explore the usefulness of both views. Building on data from French, we first show that distributional differences between average difference vectors capture semantic similarity across derivational processes in a manner comparable to the expectations of expert morphologists. We then propose an operational implementation of the classical view of rivalry based on computational classifiers: processes are rivals if and only if a classifier is unable to discriminate between them. Experimentation with French data shows that this operationalization correctly captures the broad brushes of rivalry, but also reveals finer, gradient aspects of competition in the spirit of gradient rivalry. [ABSTRACT FROM AUTHOR]

The human body is perhaps the most ubiquitous and salient visual stimulus that we encounter in our daily lives. Given the prevalence of images of human bodies in natural scene statistics, it is no surprise that our mental representations of the body are thought to strongly originate from visual experience. Yet, little is still known about high-level cognitive representations of the body. Here, we retrieved a body map from natural language, taking this as a window into high-level cognitive processes. We first extracted a matrix of distances between body parts from natural language data and employed this matrix to extrapolate a body map. To test the effectiveness of this high-level body map, we then conducted a series of experiments in which participants were asked to classify the distance between pairs of body parts, presented either as words or images. We found that the high-level body map was systematically activated when participants were making these distance judgments. Crucially, the linguistic map explained participants’ performance over and above the visual body map, indicating that the former cannot be simply conceived as a by-product of perceptual experience. These findings, therefore, establish the existence of a behaviorally relevant, high-level representation of the human body.

There is ongoing discussion about how to conceptualize the nature of the distinction between inflection and derivation. A common approach relies on qualitative differences in the semantic relationship between inflectionally versus derivationally related words: inflection yields ways to discuss the same concept in different syntactic contexts, while derivation gives rise to words for related concepts. This differential can be expected to manifest in the predictability of word frequency between words that are related derivationally or inflectionally: predicting the token frequency of a word based on information about its base form or about related words should be easier when the two words are in an inflectional relationship, rather than a derivational one. We compare prediction error magnitude for statistical models of token frequency based on distributional and frequency information of inflectionally or derivationally related words in French. The results conform to expectations: it is easier to predict the frequency of a word from properties of an inflectionally related word than from those of a derivationally related word. Prediction error provides a quantitative, continuous method to explore differences between individual processes and differences yielded by employing different predicting information, which in turn can be used to draw conclusions about the nature and manifestation of the inflection–derivation distinction.

This study uses word embeddings to investigate the semantic changes underlying the creation of two adversative connectives in Portuguese, porém and mas 'but, however'. For porém, we chart its development from an original PP formed by a preposition with a causal meaning (por) and a demonstrative pronoun that referred anaphorically to a previous proposition (en(de)). For mas, we trace its change from an adverb meaning 'more'. Adopting a distributional semantics approach, we use word embedding models trained on two corpora, the CIPM (Corpus Informatizado do Português Medieval, containing texts from the 12th–16th centuries) and COLONIA (containing texts from the 16th–20th centuries). We produce a measure of change based on the similarity scores of porém and mas with respect to words in relevant semantic categories in each corpus, representing the source and the target meanings. This paper, which constitutes the first computational study of semantic change in Portuguese, also discusses challenges and outlines steps to be taken into consideration when choosing embedding algorithms for small historical corpora. [ABSTRACT FROM AUTHOR]

This paper provides a comparative analysis of cross-lingual word embedding by studying the impact of different variables on the quality of the embedding models within the distributional semantics framework. Distributional semantics is a method for the semantic representation of words, phrases, sentences, and documents. This method aims at capturing as much information as possible from the contextual information in a vector space. The early study in this domain focused on monolingual word embedding. Further progress used cross-lingual data to capture the contextual semantic information across different languages. The main contribution of this research is to make a comparative study to find out the superior impact of the learning methods, supervised and unsupervised in training and post-training approaches in different embedding algorithms, to capture semantic properties of the words in cross-lingual embedding models to be applicable in tasks that deal with multi-languages, such as question retrieval. To this end, we study the cross-lingual embedding models created by BilBOWA, VecMap, and MUSE embedding algorithms along with the variables that impact the embedding models' quality, namely the size of the training data and the window size of the local context. In our study, we use the unsupervised monolingual Word2Vec embedding model as the baseline and evaluate the quality of embeddings on three data sets: Google analogy, mono- and cross-lingual words similar lists. We further investigated the impact of the embedding models in the question retrieval task. [ABSTRACT FROM AUTHOR]

Ambiguity in word meanings is a long-standing challenge in processing natural language. Word sense disambiguation (WSD) deals with this challenge. Prior neural language models make use of recurrent neural network and architecture with long short-term memory. These models process the words in sequence, are slower and not truly bi-directional, so they are not able to capture and represent the contextual meanings of the words, hence they are not competent in contextual semantic representation for WSD. Recent, Bi-Directional Encoder Representation from Transformers (BERT) is long short-term memory-based transformer model that is deeply bi-directional. It uses attention mechanisms, which process and use the relevance of the entire context at a time in both directions, so it is well suited to leverage the meanings in distributed representation for WSD. We have used BERT for obtaining contextual word embedding of context and sense gloss of Marathi language ambiguous word. For this purpose, we have used 282 moderately ambiguous Marathi words catering to 1004 senses distributed over 5282 Marathi sentences harvested by linguists from online Marathi websites. We have calculated semantic similarity between the pair of context and gloss embedding using Minkowski distance family and cosine similarity measures and assigned plausible sense to the given Marathi ambiguous word. Our empirical evaluation shows that the cosine similarity measure outperforms and yields an average disambiguation accuracy of 75.26% for the given Marathi sentence. [ABSTRACT FROM AUTHOR]

The distributional pattern of words in language forms the basis of linguistic distributional knowledge and contributes to conceptual processing, yet many questions remain regarding its role in cognition. We propose that corpus-based linguistic distributional models can represent a cognitively plausible approach to understanding linguistic distributional knowledge when assumed to represent an essential component of semantics, when trained on corpora representative of human language experience, and when they capture the diverse distributional relations that are useful to cognition. Using an extensive set of cognitive tasks that vary in the complexity of conceptual processing required, we systematically evaluate a wide range of model families, corpora, and parameters, and demonstrate that there is no one-size-fits-all approach for how linguistic distributional knowledge is used across cognition. Rather, linguistic distributional knowledge is a rich source of information about the world that can be accessed flexibly according to the conceptual complexity of the task at hand. [ABSTRACT FROM AUTHOR]

Distributional semantics has deeply changed in the last decades. First, predict models stole the thunder from traditional count ones, and more recently both of them were replaced in many NLP applications by contextualized vectors produced by neural language models. Although an extensive body of research has been devoted to Distributional Semantic Model (DSM) evaluation, we still lack a thorough comparison with respect to tested models, semantic tasks, and benchmark datasets. Moreover, previous work has mostly focused on task-driven evaluation, instead of exploring the differences between the way models represent the lexical semantic space. In this paper, we perform a large-scale evaluation of type distributional vectors, either produced by static DSMs or obtained by averaging the contextualized vectors generated by BERT. First of all, we investigate the performance of embeddings in several semantic tasks, carrying out an in-depth statistical analysis to identify the major factors influencing the behavior of DSMs. The results show that (i) the alleged superiority of predict based models is more apparent than real, and surely not ubiquitous and (ii) static DSMs surpass BERT representations in most out-of-context semantic tasks and datasets. Furthermore, we borrow from cognitive neuroscience the methodology of Representational Similarity Analysis (RSA) to inspect the semantic spaces generated by distributional models. RSA reveals important differences related to the frequency and part-of-speech of lexical items. [ABSTRACT FROM AUTHOR]

There is ongoing discussion about how to conceptualize the nature of the distinction between inflection and derivation. A common approach relies on qualitative differences in the semantic relationship between inflectionally versus derivationally related words: inflection yields ways to discuss the same concept in different syntactic contexts, while derivation gives rise to words for related concepts. This differential can be expected to manifest in the predictability of word frequency between words that are related derivationally or inflectionally: predicting the token frequency of a word based on information about its base form or about related words should be easier when the two words are in an inflectional relationship, rather than a derivational one. We compare prediction error magnitude for statistical models of token frequency based on distributional and frequency information of inflectionally or derivationally related words in French. The results conform to expectations: it is easier to predict the frequency of a word from properties of an inflectionally related word than from those of a derivationally related word. Prediction error provides a quantitative, continuous method to explore differences between individual processes and differences yielded by employing different predicting information, which in turn can be used to draw conclusions about the nature and manifestation of the inflection–derivation distinction. [ABSTRACT FROM AUTHOR]

Copyright of Revista Signos is the property of Pontificia Universidad Catolica de Valparaiso and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This article offers an exploratory quantitative analysis of the conceptual career of climate in US English over the period 1800–2010. Our aim is to qualify two, closely related arguments circulating in Environmental Humanities scholarship regarding the concept's history, namely that we only started to think of climate as a global entity aft er the introduction of general circulation models during the final quarter of the twentieth century, and, second, that climatic change only became an issue of environmental concern once scientists began to approach climate as a global model. While we do not dispute that the computer revolution resulted in a significantly new understanding of climate, our analysis points to a longer process of singularization and growing abstraction starting in the early nineteenth century that might help to nuance and deepen insights developed in environmental history. [ABSTRACT FROM AUTHOR]

The emergence of key concepts in Reinhart Koselleck's sense has been much discussed in conceptual history, but mainly for the eighteenth and nineteenth centuries. The article documents a post–World War II emergence of the concept of health, from relative anonymity to becoming a key concept, comparable to concepts such as politics, democracy, and culture. While previous research has emphasized conceptual mobility, this article focuses on conceptual aggregation, where the concept of health assembles and assimilates meanings, becoming essential to discourse. This is explained with reference to the development of the welfare state and the political use of a positive, expanded health concept. The article utilizes a collocation analysis of Norwegian digitized newspapers 1950–2010, culled from the uniquely extensive database of the Norwegian National Library. [ABSTRACT FROM AUTHOR]