Extrapolating human judgments from skip-gram vector representations of word meaning

Statistical Relationships Between Phonological Form, Emotional Valence and Arousal of Spanish Words

A number of studies have provided evidence of limited non-arbitrary associations between the phonological forms and meanings of affective words, a finding referred to as affective sound symbolism. Here, we explored whether the affective connotations of Spanish words might have more extensive statistical relationships with phonological/phonetic features, or affective form typicality. After eliminating words with poor affective rating agreement and morphophonological redundancies (e.g., negating prefixes), we found evidence of significant form typicality for emotional valence, emotionality, and arousal in a large sample of monosyllabic and polysyllabic words. These affective form-meaning mappings remained significant even when controlling for a range of lexico-semantic variables. We show that affective variables and their corresponding form typicality measures are able to significantly predict lexical decision performance using a megastudy dataset. Overall, our findings provide new evidence that affective form typicality is a statistical property of the Spanish lexicon.

Valence without meaning: Investigating form and semantic components in pseudowords valence

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.

How cognitive selection affects language change

Like biological species, words in language must compete to survive. Previously, it has been shown that language changes in response to cognitive constraints and over time becomes more learnable. Here, we use two complementary research paradigms to demonstrate how the survival of existing word forms can be predicted by psycholinguistic properties that impact language production. In the first study, we analyzed the survival of words in the context of interpersonal communication. We analyzed data from a large-scale serial-reproduction experiment in which stories were passed down along a transmission chain over multiple participants. The results show that words that are acquired earlier in life, more concrete, more arousing, and more emotional are more likely to survive retellings. We reason that the same trend might scale up to language evolution over multiple generations of natural language users. If that is the case, the same set of psycholinguistic properties should also account for the change of word frequency in natural language corpora over historical time. That is what we found in two large historical-language corpora (Study 2): Early acquisition, concreteness, and high arousal all predict increasing word frequency over the past 200 y. However, the two studies diverge with respect to the impact of word valence and word length, which we take up in the discussion. By bridging micro-level behavioral preferences and macro-level language patterns, our investigation sheds light on the cognitive mechanisms underlying word competition.

Semantic gender: Norms for 24,000 Dutch words and its role in word meaning

Semantic gender norms are presented for 24,037 Dutch words. Eighty participants rated 6017 words each on a five-point Likert scale ranging from feminine to masculine. Each word was rated by ten male and ten female participants. The collected norms show high reliability and correlate well with similar norms in English. We show that semantic gender is distinct from other lexical dimensions such as valence, arousal, dominance, concreteness, and age of acquisition. Semantic gender is not the same as the grammatical gender of words, either. The collected norms can be predicted accurately using a semantic space based on word association data. A dimension explaining a good amount of variance is present in this space, indicating that semantic gender is an important component of the human meaning system.

The good, the bad, and the ambivalent: Extrapolating affective values for 38,000+ Chinese words via a computational model

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.

Ratings of valence, arousal, happiness, anger, fear, sadness, disgust, and surprise for 24,000 Dutch words

Emotion is a fundamental aspect of human life and therefore is critically encoded in language. To facilitate research into the encoding of emotion in language and how emotion associations affect language processing, we present a new set of emotion norms for over 24,000 Dutch words. The emotion norms include ratings of two key dimensions of emotion: valence and arousal, as well as ratings on discrete emotion categories: happiness, anger, fear, sadness, disgust, and surprise. We show that emotional information can predict word processing, such that responses to positive words are facilitated in contrast to neutral and negative words. We also demonstrate how the ratings of emotion are related to personality characteristics. The data are available via the Open Science Framework ( https://osf.io/9htuv/ ) and serve as a valuable resource for research into emotion as well as in applied settings such as healthcare and digital communication.

Why are human animacy judgments continuous rather than categorical? A computational modeling approach

Introduction

The concept of animacy is often taken as a basic natural concept, in part I because most cases seem unambiguous. Most entities either are or are not animate. However, human animacy judgments do not reflect this binary classification. They suggest that there are borderline cases, such as virus, amoeba, fly, and imaginary beings (giant, dragon, god). Moreover, human roles (professor, mother, girlfriend) are consistently recognized as animate by far less than 100% of human judges.

Method

In this paper, I use computational modeling to identify features associated with human animacy judgments, modeling human animacy and living/non-living judgments using both bottom-up predictors (the principal components from a word embedding model) and top-down predictors (cosine distances from the names of animate categories).

Results

The results suggest that human animacy judgments may be relying on information obtained from imperfect estimates of category membership that are reflected in the word embedding models. Models using cosine distance from category names mirror human judgments in distinguishing strongly between humans (estimated lower animacy by the measure) and other animals (estimated higher animacy by the measure).

Discussion

These results are consistent with a family resemblance approach to the apparently categorical concept of animacy.

Diachronic semantic change in language is constrained by how people use and learn language

While it has long been understood that the human mind evolved to learn language, recent studies have begun to ask the inverted question: How has language evolved under the cognitive constraints of its users and become more learnable over time? In this paper, we explored how the semantic change of English words is shaped by the way humans acquire and process language. In Study 1, we quantified the extent of semantic change over the past 200 years and found that meaning change is more likely for words that are acquired later in life and are more difficult to process. We argue that it is human cognition that constrains the semantic evolution of words, rather than the other way around, because historical meanings of words were not easily accessible to people living today, and therefore could not have directly influenced how they learn and process language. In Study 2, we went further to show that semantic change, while bringing the benefit of meeting communicative needs, is cognitively costly for those who were born early enough to experience the change: Semantic change between 1970 and 2000 hindered processing speeds among middle-aged adults (ages 45–55) but not in younger adults (ages <25) in a semantic decision task. This hampering effect may have, in turn, curbed the rate of semantic change so that language does not change too fast for the human mind to catch up. Taken together, our research demonstrates that semantic change is shaped by processing and acquisition patterns across generations of language users.

The Development of Videoconference-Based Support for People Living With Rare Dementias and Their Carers: Protocol for a 3-Phase Support Group Evaluation

Background

People living with rarer dementias face considerable difficulty accessing tailored information, advice, and peer and professional support. Web-based meeting platforms offer a critical opportunity to connect with others through shared lived experiences, even if they are geographically dispersed, particularly during the COVID-19 pandemic.

Objective

We aim to develop facilitated videoconferencing support groups (VSGs) tailored to people living with or caring for someone with familial or sporadic frontotemporal dementia or young-onset Alzheimer disease, primary progressive aphasia, posterior cortical atrophy, or Lewy body dementia. This paper describes the development, coproduction, field testing, and evaluation plan for these groups.

Methods

We describe a 3-phase approach to development. First, information and knowledge were gathered as part of a coproduction process with members of the Rare Dementia Support service. This information, together with literature searches and consultation with experts by experience, clinicians, and academics, shaped the design of the VSGs and session themes. Second, field testing involved 154 Rare Dementia Support members (people living with dementia and carers) participating in 2 rounds of facilitated sessions across 7 themes (health and social care professionals, advance care planning, independence and identity, grief and loss, empowering your identity, couples, and hope and dementia). Third, a detailed evaluation plan for future rounds of VSGs was developed.

Results

The development of the small groups program yielded content and structure for 9 themed VSGs (the 7 piloted themes plus a later stages program and creativity club for implementation in rounds 3 and beyond) to be delivered over 4 to 8 sessions. The evaluation plan incorporated a range of quantitative (attendance, demographics, and geography; pre-post well-being ratings and surveys; psycholinguistic analysis of conversation; facial emotion recognition; facilitator ratings; and economic analysis of program delivery) and qualitative (content and thematic analysis) approaches. Pilot data from round 2 groups on the pre-post 3-word surveys indicated an increase in the emotional valence of words selected after the sessions.

Conclusions

The involvement of people with lived experience of a rare dementia was critical to the design, development, and delivery of the small virtual support group program, and evaluation of this program will yield convergent data about the impact of tailored support delivered to geographically dispersed communities. This is the first study to design and plan an evaluation of VSGs specifically for people affected by rare dementias, including both people living with a rare dementia and their carers, and the outcome of the evaluation will be hugely beneficial in shaping specific and targeted support, which is often lacking in this population.

International Registered Report Identifier (IRRID)

DERR1-10.2196/35376

Beyond the Benchmarks: Toward Human-Like Lexical Representations

To process language in a way that is compatible with human expectations in a communicative interaction, we need computational representations of lexical properties that form the basis of human knowledge of words. In this article, we concentrate on word-level semantics. We discuss key concepts and issues that underlie the scientific understanding of the human lexicon: its richly structured semantic representations, their ready and continual adaptability, and their grounding in crosslinguistically valid conceptualization. We assess the state of the art in natural language processing (NLP) in achieving these identified properties, and suggest ways in which the language sciences can inspire new approaches to their computational instantiation.

Extrapolation of Human Estimates of the Concreteness/ Abstractness of Words by Neural Networks of Various Architectures

In a great deal of theoretical and applied cognitive and neurophysiological research, it is essential to have more vocabularies with concreteness/abstractness ratings. Since creating such dictionaries by interviewing informants is labor-intensive, considerable effort has been made to machine-extrapolate human rankings. The purpose of the article is to study the possibility of the fast construction of high-quality machine dictionaries. In this paper, state-of-the-art deep learning neural networks are involved for the first time to solve this problem. For the English language, the BERT model has achieved a record result for the quality of a machine-generated dictionary. It is known that the use of multilingual models makes it possible to transfer ratings from one language to another. However, this approach is understudied so far and the results achieved so far are rather weak. Microsoft’s Multilingual-MiniLM-L12-H384 model also obtained the best result to date in transferring ratings from one language to another. Thus, the article demonstrates the advantages of transformer-type neural networks in this task. Their use will allow the generation of good-quality dictionaries in low-resource languages. Additionally, we study the dependence of the result on the amount of initial data and the number of languages in the multilingual case. The possibilities of transferring into a certain language from one language and from several languages together are compared. The influence of the volume of training and test data has been studied. It has been found that an increase in the amount of training data in a multilingual case does not improve the result.

Automatic generation of a large dictionary with concreteness/abstractness ratings based on a small human dictionary

Concrete/abstract words are used in a growing number of psychological and neurophysiological research. For a few languages, large dictionaries have been created manually. This is a very time-consuming and costly process. To generate large high-quality dictionaries of concrete/abstract words automatically one needs extrapolating the expert assessments obtained on smaller samples. The research question that arises is how small such samples should be to do a good enough extrapolation. In this paper, we present a method for automatic ranking concreteness of words and propose an approach to significantly decrease amount of expert assessment. The method has been evaluated on a large test set for English. The quality of the constructed dictionaries is comparable to the expert ones. The correlation between predicted and expert ratings is higher comparing to the state-of-the-art methods.

Patterns in CAOSS: Distributed representations predict variation in relational interpretations for familiar and novel compound words

While distributional semantic models that represent word meanings as high-dimensional vectors induced from large text corpora have been shown to successfully predict human behavior across a wide range of tasks, they have also received criticism from different directions. These include concerns over their interpretability (how can numbers specifying abstract, latent dimensions represent meaning?) and their ability to capture variation in meaning (how can a single vector representation capture multiple different interpretations for the same expression?). Here, we demonstrate that semantic vectors can indeed rise up to these challenges, by training a mapping system (a simple linear regression) that predicts inter-individual variation in relational interpretations for compounds such as wood brush (for example brush FOR wood, or brush MADE OF wood) from (compositional) semantic vectors representing the meanings of these compounds. These predictions consistently beat different random baselines, both for familiar compounds (moon light, Experiment 1) as well as novel compounds (wood brush, Experiment 2), demonstrating that distributional semantic vectors encode variations in qualitative interpretations that can be decoded using techniques as simple as linear regression.

Computational Models of Readers' Apperceptive Mass

Recent progress in machine-learning-based distributed semantic models (DSMs) offers new ways to simulate the apperceptive mass (AM; Kintsch, 1980) of reader groups or individual readers and to predict their performance in reading-related tasks. The AM integrates the mental lexicon with world knowledge, as for example, acquired via reading books. Following pioneering work by Denhière and Lemaire (2004), here, we computed DSMs based on a representative corpus of German children and youth literature (Jacobs et al., 2020) as null models of the part of the AM that represents distributional semantic input, for readers of different reading ages (grades 1–2, 3–4, and 5–6). After a series of DSM quality tests, we evaluated the performance of these models quantitatively in various tasks to simulate the different reader groups' hypothetical semantic and syntactic skills. In a final study, we compared the models' performance with that of human adult and children readers in two rating tasks. Overall, the results show that with increasing reading age performance in practically all tasks becomes better. The approach taken in these studies reveals the limits of DSMs for simulating human AM and their potential for applications in scientific studies of literature, research in education, or developmental science.

Augmenting Semantic Lexicons Using Word Embeddings and Transfer Learning

Sentiment-aware intelligent systems are essential to a wide array of applications. These systems are driven by language models which broadly fall into two paradigms: Lexicon-based and contextual. Although recent contextual models are increasingly dominant, we still see demand for lexicon-based models because of their interpretability and ease of use. For example, lexicon-based models allow researchers to readily determine which words and phrases contribute most to a change in measured sentiment. A challenge for any lexicon-based approach is that the lexicon needs to be routinely expanded with new words and expressions. Here, we propose two models for automatic lexicon expansion. Our first model establishes a baseline employing a simple and shallow neural network initialized with pre-trained word embeddings using a non-contextual approach. Our second model improves upon our baseline, featuring a deep Transformer-based network that brings to bear word definitions to estimate their lexical polarity. Our evaluation shows that both models are able to score new words with a similar accuracy to reviewers from Amazon Mechanical Turk, but at a fraction of the cost.

Exploring What Is Encoded in Distributional Word Vectors: A Neurobiologically Motivated Analysis

The pervasive use of distributional semantic models or word embeddings for both cognitive modeling and practical application is because of their remarkable ability to represent the meanings of words. However, relatively little effort has been made to explore what types of information are encoded in distributional word vectors. Knowing the internal knowledge embedded in word vectors is important for cognitive modeling using distributional semantic models. Therefore, in this paper, we attempt to identify the knowledge encoded in word vectors by conducting a computational experiment using Binder et al.'s (2016) featural conceptual representations based on neurobiologically motivated attributes. In an experiment, these conceptual vectors are predicted from text-based word vectors using a neural network and linear transformation, and prediction performance is compared among various types of information. The analysis demonstrates that abstract information is generally predicted more accurately by word vectors than perceptual and spatiotemporal information, and specifically, the prediction accuracy of cognitive and social information is higher. Emotional information is also found to be successfully predicted for abstract words. These results indicate that language can be a major source of knowledge about abstract attributes, and they support the recent view that emphasizes the importance of language for abstract concepts. Furthermore, we show that word vectors can capture some types of perceptual and spatiotemporal information about concrete concepts and some relevant word categories. This suggests that language statistics can encode more perceptual knowledge than often expected.

Using EEG to decode semantics during an artificial language learning task

BACKGROUND

As we learn a new nonnative language (L2), we begin to build a new map of concepts onto orthographic representations. Eventually, L2 can conjure as rich a semantic representation as our native language (L1). However, the neural processes for mapping a new orthographic representation to a familiar meaning are not well understood or characterized.

METHODS

Using electroencephalography and an artificial language that maps symbols to English words, we show that it is possible to use machine learning models to detect a newly formed semantic mapping as it is acquired.

RESULTS

Through a trial-by-trial analysis, we show that we can detect when a new semantic mapping is formed. Our results show that, like word meaning representations evoked by a L1, the localization of the newly formed neural representations is highly distributed, but the representation may emerge more slowly after the onset of the symbol. Furthermore, our mapping of word meanings to symbols removes the confound of the semantics to the visual characteristics of the stimulus, a confound that has been difficult to disentangle previously.

CONCLUSION

We have shown that the L1 semantic representation conjured by a newly acquired L2 word can be detected using decoding techniques, and we give the first characterization of the emergence of that mapping. Our work opens up new possibilities for the study of semantic representations during L2 learning.

Emotional Valence Precedes Semantic Maturation of Words: A Longitudinal Computational Study of Early Verbal Emotional Anchoring

We present a longitudinal computational study on the connection between emotional and amodal word representations from a developmental perspective. In this study, children's and adult word representations were generated using the latent semantic analysis (LSA) vector space model and Word Maturity methodology. Some children's word representations were used to set a mapping function between amodal and emotional word representations with a neural network model using ratings from 9-year-old children. The neural network was trained and validated in the child semantic space. Then, the resulting neural network was tested with adult word representations using ratings from an adult data set. Samples of 1210 and 5315 words were used in the child and the adult semantic spaces, respectively. Results suggested that the emotional valence of words can be predicted from amodal vector representations even at the child stage, and accurate emotional propagation was found in the adult word vector representations. In this way, different propagative processes were observed in the adult semantic space. These findings highlight a potential mechanism for early verbal emotional anchoring. Moreover, different multiple linear regression and mixed-effect models revealed moderation effects for the performance of the longitudinal computational model. First, words with early maturation and subsequent semantic definition promoted emotional propagation. Second, an interaction effect between age of acquisition and abstractness was found to explain model performance. The theoretical and methodological implications are discussed.

Language patterns of outgroup prejudice

Although explicit verbal expression of prejudice and stereotypes may have become less common due to the recent rise of social norms against prejudice, prejudice in language still persists in more subtle forms. It remains unclear whether and how language patterns predict variance in prejudice across a large number of minority groups. Informed by construal level theory, intergroup-contact theory, and linguistic expectancy bias, we leverage a natural language corpus of 1.8 million newspaper articles to investigate patterns of language referencing 60 U.S. minority groups. We found that perception of social distance among immigrant groups is reflected in language production: Groups perceived as socially distant (vs. close) are also more likely to be mentioned in abstract (vs. concrete) language. Concreteness was also strongly positively correlated with sentiment, a phenomenon that was unique to language concerning minority groups, suggesting a strong tendency for more socially distant groups to be represented with more negative language. We also provide a qualitative exploration of the content of outgroup prejudice by applying Latent Dirichlet Allocation to language referencing minority groups in the context of immigration. We identified 15 immigrant-related topics (e.g., politics, arts, crime, illegal workers, museums, food) and the strength of their association and relationship with perceived sentiment for each minority group. This research demonstrates how perceived social distance and language concreteness are related and correlate with outgroup negativity, provides a practical and ecologically valid method for investigating perceptions of minority groups in language, and helps elaborate the connection between theoretical positions from social psychology with recent studies from computer science on prejudice embedded in natural language.

Judgment errors in naturalistic numerical estimation

People estimate numerical quantities (such as the calories of foods) on a day-to-day basis. Although these estimates influence behavior and determine wellbeing, they are prone to two important types of errors. Scaling errors occur when people make mistakes reporting their beliefs about a particular numerical quantity (e.g. by inflating small numbers). Belief errors occur when people make mistakes using their knowledge of the judgment target to form their beliefs about the numerical quantity (e.g. by overweighting certain cues). In this paper, we quantitatively model numerical estimates, and in turn, scaling and belief errors, in everyday judgment tasks. Our approach is unique in using insights from semantic memory research to specify knowledge for naturalistic judgment targets, allowing our models to formally describe nuanced errors in belief not considered in prior research. In Studies 1 and 2, we find that belief error models predict participant estimates and errors with very high out-of-sample accuracy rates, significantly outperforming the predictions of scaling error models. In fact, the best-fitting belief error models can closely mimic the inverse-S shaped patterns captured by scaling error models, suggesting that the types of responses previously attributed to scaling errors can be seen as errors of belief. In Studies 3 to 8, we find that belief error models are also able to predict people's responses in semantic judgment, free association, and verbal protocol tasks related to numerical judgment, and thus provide a good account of the cognitive underpinnings of judgment.

The Croatian psycholinguistic database: Estimates for 6000 nouns, verbs, adjectives and adverbs

Psycholinguistic databases containing ratings of concreteness, imageability, age of acquisition, and subjective frequency are used in psycholinguistic and neurolinguistic studies which require words as stimuli. Linguistic characteristics (e.g. word length, corpus frequency) are frequently coded, but word class is seldom systematically treated, although there are indications of its significance for imageability and concreteness. This paper presents the Croatian Psycholinguistic Database (CPD; available at: https://doi.org/10.17234/megahr.2019.hpb ), containing 6000 Croatian nouns, verbs, adjectives and adverbs, rated for concreteness, imageability, age of acquisition, and subjective frequency. Moreover, we present computationally obtained extrapolations of concreteness and imageability to the remainder of the Croatian lexicon (available at: https://github.com/megahr/lexicon/blob/master/predictions/hr_c_i.predictions.txt ). In the two studies presented here, we explore the significance of word class for concreteness and imageability in human and computationally obtained ratings. The observed correlations in the CPD indicate correspondences between psycholinguistic measures expected from the literature. Word classes exhibit differences in subjective frequency, age of acquisition, concreteness and imageability, with significant differences between nouns, verbs, adjectives and adverbs. In the computational study which focused on concreteness and imageability, concreteness obtained higher correlations with human ratings than imageability, and the system underpredicted the concreteness of nouns, and overpredicted the concreteness of adjectives and adverbs. Overall, this suggests that word class contains schematic conceptual and distributional information. Schematic conceptual content seems to be more significant in human ratings of concreteness and less significant in computationally obtained ratings, where distributional information seems to play a more significant role. This suggests that word class differences should be theoretically explored.

Sound symbolism shapes the English language: The maluma/takete effect in English nouns

Sound symbolism refers to associations between language sounds (i.e., phonemes) and perceptual and/or semantic features. One example is the maluma/takete effect: an association between certain phonemes (e.g., /m/, /u/) and roundness, and others (e.g., /k/, /ɪ/) and spikiness. While this association has been demonstrated in laboratory tasks with nonword stimuli, its presence in existing spoken language is unknown. Here we examined whether the maluma/takete effect is attested in English, across a broad sample of words. Best-worst judgments from 171 university students were used to quantify the shape of 1,757 objects, from spiky to round. We then examined whether the presence of certain phonemes in words predicted the shape of the objects to which they refer. We found evidence that phonemes associated with roundness are more common in words referring to round objects, and phonemes associated with spikiness are more common in words referring to spiky objects. This represents an instance of iconicity, and thus nonarbitrariness, in human language.

Words as a window: Using word embeddings to explore the learned representations of Convolutional Neural Networks

As deep neural net architectures minimize loss, they accumulate information in a hierarchy of learned representations that ultimately serve the network's final goal. Different architectures tackle this problem in slightly different ways, but all create intermediate representational spaces built to inform their final prediction. Here we show that very different neural networks trained on two very different tasks build knowledge representations that display similar underlying patterns. Namely, we show that the representational spaces of several distributional semantic models bear a remarkable resemblance to several Convolutional Neural Network (CNN) architectures (trained for image classification). We use this information to explore the network behavior of CNNs (1) in pretrained models, (2) during training, and (3) during adversarial attacks. We use these findings to motivate several applications aimed at improving future research on CNNs. Our work illustrates the power of using one model to explore another, gives new insights into the function of CNN models, and provides a framework for others to perform similar analyses when developing new architectures. We show that one neural network model can provide a window into understanding another.

Sentiment Analysis of Children and Youth Literature: Is There a Pollyanna Effect?

If the words of natural human language possess a universal positivity bias, as assumed by Boucher and Osgood’s (1969) famous Pollyanna hypothesis and computationally confirmed for large text corpora in several languages (Dodds et al., 2015), then children and youth literature (CYL) should also show a Pollyanna effect. Here we tested this prediction applying an unsupervised vector space model-based sentiment analysis tool called SentiArt (Jacobs, 2019) to two CYL corpora, one in English (372 books) and one in German (500 books). Pitching our analysis at the sentence level, and assessing semantic as well as lexico-grammatical information, both corpora show the Pollyanna effect and thus add further evidence to the universality hypothesis. The results of our multivariate sentiment analyses provide interesting testable predictions for future scientific studies of literature.

Redundancy, isomorphism, and propagative mechanisms between emotional and amodal representations of words: A computational study

Some proposals claim that language acts as a link to propagate emotional and other modal information. Thus, there is an eminently amodal path of emotional propagation in the mental lexicon. Following these proposals, we present a computational model that emulates a linking mechanism (mapping function) between emotional and amodal representations of words using vector space models, emotional feature-based models, and neural networks. We analyzed three central concepts within the embodiment debate (redundancy, isomorphism, and propagative mechanisms) comparing two alternative hypotheses: semantic neighborhood hypothesis versus specific dimensionality hypothesis. Univariate and multivariate neural networks were trained for dimensional (N = 11,357) and discrete emotions (N = 2,266), and later we analyzed its predictions in a test set (N = 4,167 and N = 875, respectively). We showed how this computational model could propagate emotional responses to words without a direct emotional experience via amodal propagation, but no direct relations were found between emotional rates and amodal distances. Thereby, we found that there were clear redundancy and propagative mechanisms, but no isomorphism should be assumed. Results suggested that it was necessary to establish complex links to go beyond amodal distances of vector spaces. In this way, although the emotional rates of semantic neighborhoods could predict the emotional rates of target words, the mapping function of specific amodal features seemed to simulate emotional responses better. Thus, both hypotheses would not be mutually exclusive. We also showed that discrete emotions could have simpler relations between modal and amodal representations than dimensional emotions. All these results and their theoretical implications are discussed.

subs2vec: Word embeddings from subtitles in 55 languages

This paper introduces a novel collection of word embeddings, numerical representations of lexical semantics, in 55 languages, trained on a large corpus of pseudo-conversational speech transcriptions from television shows and movies. The embeddings were trained on the OpenSubtitles corpus using the fastText implementation of the skipgram algorithm. Performance comparable with (and in some cases exceeding) embeddings trained on non-conversational (Wikipedia) text is reported on standard benchmark evaluation datasets. A novel evaluation method of particular relevance to psycholinguists is also introduced: prediction of experimental lexical norms in multiple languages. The models, as well as code for reproducing the models and all analyses reported in this paper (implemented as a user-friendly Python package), are freely available at: https://github.com/jvparidon/subs2vec.

Representation of associative and affective semantic similarity of abstract words in the lateral temporal perisylvian language regions

The examination of semantic cognition has traditionally identified word concreteness as well as valence as two of the principal dimensions in the representation of conceptual knowledge. More recently, corpus-based vector space models as well as graph-theoretical analysis of large-scale task-related behavioural responses have revolutionized our insight into how the meaning of words is structured. In this fMRI study, we apply representational similarity analysis to investigate the conceptual representation of abstract words. Brain activity patterns were related to a cued-association based graph as well as to a vector-based co-occurrence model of word meaning. Twenty-six subjects (19 females and 7 males) performed an overt repetition task during fMRI. First, we performed a searchlight classification procedure to identify regions where activity is discriminable between abstract and concrete words. These regions were left inferior frontal gyrus, the upper and lower bank of the superior temporal sulcus bilaterally, posterior middle temporal gyrus and left fusiform gyrus. Representational Similarity Analysis demonstrated that for abstract words, the similarity of activity patterns in the cortex surrounding the superior temporal sulcus bilaterally and in the left anterior superior temporal gyrus reflects the similarity in word meaning. These effects were strongest for semantic similarity derived from the cued association-based graph and for affective similarity derived from either of the two models. The latter effect was mainly driven by positive valence words. This research highlights the close neurobiological link between the information structure of abstract and affective word content and the similarity in activity pattern in the lateral and anterior temporal language system.

Studying language in context using the temporal generalization method

The temporal generalization method (TGM) is a data analysis technique that can be used to test if the brain's representation for particular stimuli (e.g. sounds, images) is maintained, or if it changes as a function of time (King J-R, Dehaene S. 2014 Characterizing the dynamics of mental representations: the temporal generalization method. Trends Cogn. Sci.18, 203-210. (doi:10.1016/j.tics.2014.01.002)). The TGM involves training models to predict the stimuli or condition using a time window from a recording of brain activity, and testing the resulting models at all possible time windows. This is repeated for all possible training windows to create a full matrix of accuracy for every combination of train/test window. The results of a TGM indicate when brain activity patterns are consistent (i.e. the trained model performs well even when tested on a different time window), and when they are inconsistent, allowing us to track neural representations over time. The TGM has been used to study the representation of images and sounds during a variety of tasks, but has been less readily applied to studies of language. Here, we give an overview of the method itself, discuss how the TGM has been used to analyse two studies of language in context and explore how the TGM could be applied to further our understanding of semantic composition. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.

Conceptualizing syntactic categories as semantic categories: Unifying part-of-speech identification and semantics using co-occurrence vector averaging

Co-occurrence models have been of considerable interest to psychologists because they are built on very simple functionality. This is particularly clear in the case of prediction models, such as the continuous skip-gram model introduced in Mikolov, Chen, Corrado, and Dean (2013), because these models depend on functionality closely related to the simple Rescorla-Wagner model of discriminant learning in nonhuman animals (Rescorla & Wagner, 1972), which has a rich history within psychology as a model of many animal learning processes. We replicate and extend earlier work showing that it is possible to extract accurate information about syntactic category and morphological family membership directly from patterns of word co-occurrence, and provide evidence from four experiments showing that this information predicts human reaction times and accuracy for class membership decisions.

Predicting High-Level Human Judgment Across Diverse Behavioral Domains

Recent advances in machine learning, combined with the increased availability of large natural language datasets, have made it possible to uncover semantic representations that characterize what people know about and associate with a wide range of objects and concepts. In this paper, we examine the power of word embeddings, a popular approach for uncovering semantic representations, for studying high-level human judgment. Word embeddings are typically applied to linguistic and semantic tasks, however we show that word embeddings can be used to predict complex theoretically- and practically- relevant human perceptions and evaluations in domains as diverse as social cognition, health behavior, risk perception, organizational behavior, and marketing. By learning mappings from word embeddings directly onto judgment ratings, we outperform a similarity-based baseline and perform favorably compared to common metrics of human inter-rater reliability. Word embeddings are also able to identify the concepts that are most associated with observed perceptions and evaluations, and can thus shed light on the psychological substrates of judgment. Overall, we provide new methods and insights for predicting and understanding high-level human judgment, with important applications across the social and behavioral sciences.

Bridging the theoretical gap between semantic representation models without the pressure of a ranking: some lessons learnt from LSA

In recent years, latent semantic analysis (LSA) has reached a level of maturity at which its presence is ubiquitous in technology as well as in simulation of cognitive processes. In spite of this, in recent years there has been a trend of subjecting LSA to some criticisms, usually because it is compared to other models in very specific tasks and conditions and sometimes without having good knowledge of what the semantic representation of LSA means, and without exploiting all the possibilities of which LSA is capable other than the cosine. This paper provides a critical review to clarify some of the misunderstandings regarding LSA and other space models. The historical stability of the predecessors of LSA, the representational structure of word meaning and the multiple topologies that could arise from a semantic space, the computation of similarity, the myth that LSA dimensions have no meaning, the computational and algorithm plausibility to account for meaning acquisition in LSA (in contrast to others models based on online mechanisms), the possibilities of spatial models to substantiate recent proposals, and, in general, the characteristics of classic vector models and their ease and flexibility to simulate some cognitive phenomena will be reviewed. The review highlights the similarity between LSA and other techniques and proposes using long LSA experiences in other models, especially in predicting models such as word2vec. In sum, it emphasizes the lessons that can be learned from comparing LSA-based models to other models, rather than making statements about "the best."

Word prevalence norms for 62,000 English lemmas

We present word prevalence data for 61,858 English words. Word prevalence refers to the number of people who know the word. The measure was obtained on the basis of an online crowdsourcing study involving over 220,000 people. Word prevalence data are useful for gauging the difficulty of words and, as such, for matching stimulus materials in experimental conditions or selecting stimulus materials for vocabulary tests. Word prevalence also predicts word processing times, over and above the effects of word frequency, word length, similarity to other words, and age of acquisition, in line with previous findings in the Dutch language.

Sentiment Analysis for Words and Fiction Characters From the Perspective of Computational (Neuro-)Poetics

Two computational studies provide different sentiment analyses for text segments (e.g., "fearful" passages) and figures (e.g., "Voldemort") from the Harry Potter books (Rowling, 1997, 1998, 1999, 2000, 2003, 2005, 2007) based on a novel simple tool called SentiArt. The tool uses vector space models together with theory-guided, empirically validated label lists to compute the valence of each word in a text by locating its position in a 2d emotion potential space spanned by the words of the vector space model. After testing the tool's accuracy with empirical data from a neurocognitive poetics study, it was applied to compute emotional figure and personality profiles (inspired by the so-called "big five" personality theory) for main characters from the book series. The results of comparative analyses using different machine-learning classifiers (e.g., AdaBoost, Neural Net) show that SentiArt performs very well in predicting the emotion potential of text passages. It also produces plausible predictions regarding the emotional and personality profile of fiction characters which are correctly identified on the basis of eight character features, and it achieves a good cross-validation accuracy in classifying 100 figures into "good" vs. "bad" ones. The results are discussed with regard to potential applications of SentiArt in digital literary, applied reading and neurocognitive poetics studies such as the quantification of the hybrid hero potential of figures.

Imageability and Neighborhood Density Facilitate the Age of Word Acquisition in Czech

Purpose The study examined the effects of imageability and phonological neighborhood density on the acquisition of word production in Czech, controlling for part-of-speech class, word length, and word frequency. Phonological neighborhood density is of interest because previous research has not examined highly inflected languages such as Czech. The effects of imageability on word acquisition are widely assumed, but only a few empirical studies examined such effects using child data directly. Method Data from the Czech norming study of the MacArthur-Bates Communicative Development Inventories ( Smolík, Turková, Marušincová, & Malechová, 2017 ) adaptation were used, and all nouns and action words in the data set were examined (total 359). Based on the norming sample of 493 children, the expected age of acquisition was calculated. Results A small but significant effect of neighborhood density (explaining 1.5% of unique variance) was found, as well as a robust effect of imageability (9% of unique variance). Imageability also accounted for the difference between nouns and verbs in the age of acquisition. Conclusion Imageability is a robust predictor of word age of acquisition that should be taken into account in future studies. The identifiability of the referent and the memory mechanisms are likely responsible for the strong imageability effect. Words with large phonological neighborhoods are acquired earlier, even in a language with complex inflectional morphology.

When is best-worst best? A comparison of best-worst scaling, numeric estimation, and rating scales for collection of semantic norms

Large-scale semantic norms have become both prevalent and influential in recent psycholinguistic research. However, little attention has been directed towards understanding the methodological best practices of such norm collection efforts. We compared the quality of semantic norms obtained through rating scales, numeric estimation, and a less commonly used judgment format called best-worst scaling. We found that best-worst scaling usually produces norms with higher predictive validities than other response formats, and does so requiring less data to be collected overall. We also found evidence that the various response formats may be producing qualitatively, rather than just quantitatively, different data. This raises the issue of potential response format bias, which has not been addressed by previous efforts to collect semantic norms, likely because of previous reliance on a single type of response format for a single type of semantic judgment. We have made available software for creating best-worst stimuli and scoring best-worst data. We also made available new norms for age of acquisition, valence, arousal, and concreteness collected using best-worst scaling. These norms include entries for 1,040 words, of which 1,034 are also contained in the ANEW norms (Bradley & Lang, Affective norms for English words (ANEW): Instruction manual and affective ratings (pp. 1-45). Technical report C-1, the center for research in psychophysiology, University of Florida, 1999).

Sliding into happiness: A new tool for measuring affective responses to words

Reliable measurement of affective responses is critical for research into human emotion. Affective evaluation of words is most commonly gauged on multiple dimensions-including valence (positivity) and arousal-using a rating scale. Despite its popularity, this scale is open to criticism: It generates ordinal data that is often misinterpreted as interval, it does not provide the fine resolution that is essential by recent theoretical accounts of emotion, and its extremes may not be properly calibrated. In 5 experiments, the authors introduce a new slider tool for affective evaluation of words on a continuous, well-calibrated and high-resolution scale. In Experiment 1, participants were shown a word and asked to move a manikin representing themselves closer to or farther away from the word. The manikin's distance from the word strongly correlated with the word's valence. In Experiment 2, individual differences in shyness and sociability elicited reliable differences in distance from the words. Experiment 3 validated the results of Experiments 1 and 2 using a demographically more diverse population of responders. Finally, Experiment 4 (along with Experiment 2) suggested that task demand is not a potential cause for scale recalibration. In Experiment 5, men and women placed a manikin closer or farther from words that showed sex differences in valence, highlighting the sensitivity of this measure to group differences. These findings shed a new light on interactions among affect, language, and individual differences, and demonstrate the utility of a new tool for measuring word affect. (PsycINFO Database Record