No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
The map trap: Why and how word learning research should move beyond mapping
Abstract
A pervasive goal in the study of how children learn word meanings is to explain how young children solve the mapping problem. The mapping problem asks how language learners connect a label to its referent. Mapping is one part of word learning, however, it does not reflect other critical components of word meaning construction, such as the encoding of lexico‐semantic relations and socio‐pragmatic context. In this paper, we argue that word learning researchers' overemphasis of mapping has constrained our experimental paradigms and hypotheses, leading to misconceived theories and policy interventions. We first explain how the mapping focus limits our ability to study the richness and complexity of what infants and children learn about, and do with, word meanings. Then, we describe how our focus on mapping has constrained theory development. Specifically, we show how it has led to (a) the misguided emphasis on referent selection and ostensive labeling, and (b) the undervaluing of diverse pathways to word knowledge, both within and across cultures. We also review the consequences of the mapping focus outside of the lab, including myopic language learning interventions. Last, we outline an alternative, more inclusive approach to experimental study and theory construction in word learning research. This article is categorized under: Psychology > Language Psychology > Theory and Methods Psychology > Learning We review how despite advances in our understanding of early word representations, there remains an overemphasis on the mapping problem. After outlining the consquences of the mapping focus, we also discuss more inclusive approaches to the study of word learning.
... Given the focus on individual and grounded learning experience, such approaches appear better equipped to explain the already attested individual variation in language competence. Also, given the multiplicity of factors involved in word learning and the variation in contexts in which words are acquired, more sophisticated accounts are, in addition, moving away from classical sound-to-referent mapping approaches (Pace et al., 2016;Rohlfing et al., 2016;Wojcik et al., 2022). Current embodiment theories of cognition and language propose that human cognition and language are grounded in experience, and that concepts and their linguistic labels are formed in rich interaction with the world. ...
Word learning requires successful pairing of form and meaning. A common hypothesis about the process of word learning is that initially, infants work on identifying the phonological segments corresponding to words (speech analysis), and subsequently map those segments onto meaning. A range of theories have been proposed to account for the underlying mechanisms and factors in this remarkable achievement. While some are mainly concerned with the sensorimotor affordances and perceptual properties of referents out in the world, other theories emphasize the importance of language as a system, and the relations among language units (other words or syntax). Recent approaches inspired by neuro-science suggest that the storage and processing of word meanings is supported by neural systems subserving both the representation of conceptual knowledge and its access and use (Lambon Ralph et al., Nature Reviews Neuroscience 18:42–55, 2017). Developmental disorders have been attested to impact on different aspects of word learning. While impaired word knowledge is not a hallmark of Autism Spectrum Disorder (ASD), and remains largely understudied in this population, there is evidence that there are, sometimes subtle, problems in that domain, reflected in both how such knowledge is acquired and how words are used (Vulchanova et al., Word knowledge and word usage: A cross-disciplinary guide to the mental lexicon, Mouton De Gruyter, 2020). In addition, experimental evidence suggests that children with autism present with specific problems in categorizing the referents of linguistic labels leading to subsequent problems with using those labels (Hartley and Allen, Autism 19:570–579, 2015). Furthermore, deficits have been reported in some of the underlying mechanisms, biases and use of cues in word learning, such as e.g., object shape (Field et al., Journal of Autism and Developmental Disorders 46:1210–1219, 2016; Tek et al., Autism Research 1:208–222, 2008). Finally, it is likely that symbol use might be impaired in ASD, however, the direction of the causal relationship between social and communication impairment in autism and symbolic skills is still an open question (Allen and Lewis, Journal of Autism and Developmental Disorders 45:1–3, 2015; Allen and Butler, British Journal of Developmental Psychology 38:345–362, 2020; Wainwright et al., Journal of Autism and Developmental Disorders 50:2941–2956, 2020). Further support for impaired symbol formation in autism comes from the well-attested problems with figurative, non-literal language use (e.g., metaphors, idioms, hyperbole, irony) (Vulchanova et al., Frontiers in Human Neuroscience 9:24, 2015). Here we propose that embodied theories of cognition which link perceptual experience with conceptual knowledge (see Eigsti, Frontiers in Psychology 4:224, 2013; Klin et al., Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 358:345–360, 2003) might be useful in explaining the difficulty in symbolic understanding that individuals with autism face during the word learning process.
... We examine multiple aspects of word learning, including how well children link words to referents, encode semantic category information, encode phonological representations of the word forms, and retain these phonological representations over a delay. Our use of multiple measures provides a more holistic exploration of word learning, moving beyond the tendency to focus solely on how children link words to referents (Wojcik et al., 2022). We focus on the early stages of word learning-what information children are able to encode after only a few exposures, but not how this process begins (triggering; e.g., Hoover et al., 2010) or how it extends over time as children learn to approximate the full meaning of a word (e.g., Carey, 2010). ...
Purpose:
The current study compared the effects of direct instruction versus indirect exposure on multiple aspects of novel word learning for children with developmental language disorder (DLD) and children with typical language development (TLD).
Method:
Participants included 36 children with DLD and 45 children with TLD. All children were in the first grade and 6-8 years of age (Mdn = 7 years; 2 months). Using a between-subjects design, children were randomly assigned to be exposed to novel words and their unfamiliar referents via either direct instruction (each referent presented in isolation with an explicit goal of learning) or indirect exposure (multiple referents presented with the goal of answering yes/no questions).
Results:
In alternative forced-choice measures of recognition, children with DLD were less accurate than their TLD peers in linking words to referents, encoding semantic categories for words, and encoding detailed representations of word forms. These differences in word learning were accounted for by a constellation of cognitive measures, including receptive vocabulary, phonological memory, visuospatial memory, and sustained attention. All children were similarly accurate in retaining word forms over a 24- to 48-hr delay. Children with TLD were more accurate in all aspects of word learning following direct instruction compared to indirect exposure. Benefits from direct instruction were observed for children with DLD in link and semantic, but not word form, learning.
Conclusions:
These results suggest that vocabulary interventions with direct instruction can help children with DLD learn some, but not all, aspects of novel words. Additional support is necessary to help children with DLD encode rich phonological representations.
Children learn words in a social environment, facilitated in part by social cues from caregivers, such as eye‐gaze and gesture. A common assumption is that social cues convey either perceptual or social information, depending on the age of the child. In this review of research on word learning and social cues during early childhood, we propose that (1) the functions of social cues are not categorically perceptual or social, and (2) social cues support word learning in four interdependent ways: by helping children to orient attention, extract relevant information, disambiguate referents, and understand others' referential intent. We conclude with specific recommendations for theory‐building and suggest that the dynamic and complex functions of social cues need to be accounted for in any complete theory of word learning.
Social cues, such as eye gaze and referential gestures, have been shown to promote word learning in children, usually by exposing children to a label-object association with vs. without a social cue. However, social cues are not always so cleanly linked to referents in the world. Instead, label-object associations occur imperfectly across time in a discourse. We conducted an experiment to examine how labels and eye gaze interactively support 4- and 5-year-olds’ word learning during exposures to a series of brief discourse topics. We found that children learned novel words better when social cues and word labels jointly marked the beginning or end of a discourse topic, but not when they were misaligned. The findings indicate that young learners can use social cues to inform their understanding of word meanings across successive utterances, not just within single labeling events. We discuss how social cues may support children’s ability to find where discourse topics begin and end within their ongoing speech input.
For human children as well as machine learning systems, a key challenge in learning a word is linking the word to the visual phenomena it describes. We explore this aspect of word learning by using the performance of computer vision systems as a proxy for the difficulty of learning a word from visual cues. We show that the age at which children acquire different categories of words is predicted by the performance of visual classification and captioning systems, over and above the expected effects of word frequency. The performance of the computer vision systems is related to human judgments of the concreteness of words, supporting the idea that we are capturing the relationship between words and visual phenomena.
Infants enculturate to their soundscape over the first year of life, yet theories of how they do so rarely make contact with details about the sounds available in everyday life. Here, we report on properties of a ubiquitous early ecology in which foundational skills get built: music. We captured daylong recordings from 35 infants ages 6–12 months at home and fully double‐coded 467 h of everyday sounds for music and its features, tunes, and voices. Analyses of this first‐of‐its‐kind corpus revealed two distributional properties of infants’ everyday musical ecology. First, infants encountered vocal music in over half, and instrumental in over three‐quarters, of everyday music. Live sources generated one‐third, and recorded sources three‐quarters, of everyday music. Second, infants did not encounter each individual tune and voice in their day equally often. Instead, the most available identity cumulated to many more seconds of the day than would be expected under a uniform distribution. These properties of everyday music in human infancy are different from what is discoverable in environments highly constrained by context (e.g., laboratories) and time (e.g., minutes rather than hours). Together with recent insights about the everyday motor, language, and visual ecologies of infancy, these findings reinforce an emerging priority to build theories of development that address the opportunities and challenges of real input encountered by real learners.
Toddlers vary widely in the rate at which they develop vocabulary. This variation predicts later language development and school success at the group level; however, we cannot determine which children with slower vocabulary development in the second year will continue to have difficulty. In this article, I argue that this is because we lack theoretical understanding of how multiple processes operate as a system to create individual children’s pathways to word learning. I discuss the difficulties children face when learning even a single concrete noun, the multiple general cognitive processes that support word learning, and some evidence of rapid development in the second year. I present work toward a formal model of the word learning system and how this system changes over time. The long‐term goal of this work is to understand how individual children’s strengths and weaknesses create unique vocabulary pathways, and to enable us to predict outcomes and identify effective interventions.
We introduce a new resource: the SAYCam corpus. Infants aged 6–32 months wore a head-mounted camera for approximately 2 hr per week, over the course of approximately two-and-a-half years. The result is a large, naturalistic, longitudinal dataset of infant- and child-perspective videos. Over 200,000 words of naturalistic speech have already been transcribed. Similarly, the dataset is searchable using a number of criteria (e.g., age of participant, location, setting, objects present). The resulting dataset will be of broad use to psychologists, linguists, and computer scientists.
As adults, we draw upon our ample knowledge about the world to support such vital cognitive feats as using language, reasoning, retrieving knowledge relevant to our current goals, planning for the future, adapting to unexpected events, and navigating through the environment. Our knowledge readily supports these feats because it is not merely a collection of stored facts, but rather functions as an organized, semantic network of concepts connected by meaningful relations. How do the relations that fundamentally organize semantic concepts emerge with development? Here, we cast a spotlight on a potentially powerful but often overlooked driver of semantic organization: Rich statistical regularities that are ubiquitous in both language and visual input. In this synthetic review, we show that a driving role for statistical regularities is convergently supported by evidence from diverse fields, including computational modeling, statistical learning, and semantic development. Finally, we identify a number of key avenues of future research into how statistical regularities may drive the development of semantic organization. Keywords: semantic development; semantic organization; semantic knowledge; statistical learning; taxonomic relations; association
An important question in early bilingual first language acquisition concerns the development of lexical-semantic associations within and across two languages. The present study investigates the earliest emergence of lexical-semantic priming at 18 and 24 months in Spanish-English bilinguals (N = 32) and its relation to vocabulary knowledge within and across languages. Results indicate a remarkably similar pattern of development between monolingual and bilingual children, such that lexical-semantic development begins at 18 months and strengthens by 24 months. Further, measures of cross-language lexical knowledge are stronger predictors of children’s lexical-semantic processing skill than measures that capture single-language knowledge only. This suggests that children make use of both languages when processing semantic information. Together these findings inform the understanding of the relation between lexical-semantic breadth and organization in the context of dual language learners in early development.
Across multiple situations, learners are sensitive to co-occurrences between individual words and their referents in the environment, which provides a means by which the ambiguity of word- world mappings may be resolved. In three studies, we tested whether cross-situational learning is sufficiently powerful to support simultaneous learning the referents for words from multiple grammatical categories, a more realistic reflection of more complex natural language learning situations. In Experiment 1, learners heard sentences comprising nouns, verbs, adjectives, and grammatical markers indicating subject and object roles, and viewed a dynamic scene to which the sentence referred. In Experiments 2 and 3, we further increased the uncertainty of the referents by presenting two scenes alongside each sentence. In all studies, we found that cross- situational statistical learning was sufficiently powerful to facilitate acquisition of both vocabulary and grammar from complex sentence-to-scene correspondences, simulating the situations that more closely resemble the challenge facing the language learner.
If the structure of language vocabularies mirrors the structure of natural divisions that are universally perceived, then the meanings of words in different languages should closely align. By contrast, if shared word meanings are a product of shared culture, history and geography, they may differ between languages in substantial but predictable ways. Here, we analysed the semantic neighbourhoods of 1,010 meanings in 41 languages. The most-aligned words were from semantic domains with high internal structure (number, quantity and kinship). Words denoting natural kinds, common actions and artefacts aligned much less well. Languages that are more geographically proximate, more historically related and/or spoken by more-similar cultures had more aligned word meanings. These results provide evidence that the meanings of common words vary in ways that reflect the culture, history and geography of their users. A comparison of 41 languages reveals that words for common actions, artefacts and natural kinds are less translatable than expected. Translatability is related to the cultural similarity of language communities and to their historic relationships.
Concept induction requires the extraction and naming of concepts from noisy perceptual experience. For supervised approaches, as the number of concepts grows, so does the number of required training examples. Philosophers, psychologists and computer scientists have long recognized that children can learn to label objects without being explicitly taught. In a series of computational experiments, we highlight how information in the environment can be used to build and align conceptual systems. Unlike supervised learning, the learning problem becomes easier the more concepts and systems there are to master. The key insight is that each concept has a unique signature within one conceptual system (for example, images) that is recapitulated in other systems (for example, text or audio). As predicted, children’s early concepts form readily aligned systems. By assembling conceptual systems from real-word datasets of text, images and audio, Roads and Love propose that objects embedded within a conceptual system have a unique signature that allows for conceptual systems to be aligned in an unsupervised fashion.
Reading stories to children fosters their language development. An approach rarely investigated is narrators telling stories without reading from text (i.e., oral storytelling). Oral storytelling may differ from more commonly employed read-aloud approaches in terms of language complexity and the opportunity to regulate the storytelling process via attention-guiding behavior, such as eye contact and gesticulation. By experimentally separating the influences of language complexity and attention-guiding behavior, the current study tried to shed light on the effect of story-delivery method (oral storytelling vs. read-aloud) and its underlying mechanisms on novel word acquisition, story comprehension, and children's on-task behavior. In a 4 × 2 mixed-design, with method of story delivery (live read-aloud vs. live oral storytelling vs. audiotaped read-aloud vs. audiotaped oral storytelling) as a between-subjects factor and time (pretest vs. posttest) as a within-subjects factor, a sample of 60 four-to six-year-old children listened to four short stories in one of the four conditions twice. Target-word learning from pre-to posttest as well as story comprehension were measured. Additionally, in the live conditions storyteller and child behavior was coded. Although learning occurred across conditions, live oral storytelling resulted in the largest gains in receptive target-vocabulary and best story comprehension. In addition, children were less restless and more attentive during live oral storytelling.
Daylong at‐home audio recordings from 10 Tseltal Mayan children (0;2–3;0; Southern Mexico) were analyzed for how often children engaged in verbal interaction with others and whether their speech environment changed with age, time of day, household size, and number of speakers present. Children were infrequently directly spoken to, with most directed speech coming from adults, and no increase with age. Most directed speech came in the mornings, and interactional peaks contained nearly four times the baseline rate of directed speech. Coarse indicators of children’s language development (babbling, first words, first word combinations) suggest that Tseltal children manage to extract the linguistic information they need despite minimal directed speech. Multiple proposals for how they might do so are discussed.
The adult lexicon links concepts and labels with related meanings (e.g. dog–cat). We asked how children’s encounters with concepts versus labels contribute to their semantic development. We investigated semantic priming in monolinguals and bilinguals, who have similar experience with concepts, but different experience with labels (i.e. monolinguals hear “dog”, bilinguals hear “dog” and “chien”). Similarities in performance across monolinguals and bilinguals, as well as across bilinguals’ two languages, support the position that encounters with concepts – as opposed to labels – are the largest driver of early semantic development. Our findings also suggest that the effects of semantic priming may be challenging to observe in the current paradigm at 24 months, but are strong in bilinguals by age 30 months.
This article provides an estimation of how frequently, and from whom, children aged 0–11 years (Ns between 9 and 24) receive one-on-one verbal input among Tsimane forager-horticulturalists of lowland Bolivia. Analyses of systematic daytime behavioral observations reveal < 1 min per daylight hour is spent talking to children younger than 4 years of age, which is 4 times less than estimates for others present at the same time and place. Adults provide a majority of the input at 0–3 years of age but not afterward. When integrated with previous work, these results reveal large cross-cultural variation in the linguistic experiences provided to young children. Consideration of more diverse human populations is necessary to build generalizable theories of language acquisition.
Previous research has suggested that distributional learning mechanisms may contribute to the acquisition of semantic knowledge. However, distributional learning mechanisms, statistical learning, and contemporary “deep learning” approaches have been criticized for being incapable of learning the kind of abstract and structured knowledge that many think is required for acquisition of semantic knowledge. In this paper, we show that recurrent neural networks, trained on noisy naturalistic speech to children, do in fact learn what appears to be abstract and structured knowledge. We trained two types of recurrent neural networks (Simple Recurrent Network, and Long Short-Term Memory) to predict word sequences in a 5-million-word corpus of speech directed to children ages 0–3 years old, and assessed what semantic knowledge they acquired. We found that learned internal representations are encoding various abstract grammatical and semantic features that are useful for predicting word sequences. Assessing the organization of semantic knowledge in terms of the similarity structure, we found evidence of emergent categorical and hierarchical structure in both models. We found that the Long Short-term Memory (LSTM) and SRN are both learning very similar kinds of representations, but the LSTM achieved higher levels of performance on a quantitative evaluation. We also trained a non-recurrent neural network, Skip-gram, on the same input to compare our results to the state-of-the-art in machine learning. We found that Skip-gram achieves relatively similar performance to the LSTM, but is representing words more in terms of thematic compared to taxonomic relations, and we provide reasons why this might be the case. Our findings show that a learning system that derives abstract, distributed representations for the purpose of predicting sequential dependencies in naturalistic language may provide insight into emergence of many properties of the developing semantic system.
Previous studies have shown bilingually and monolingually developing children to differ in their sensitivity to referential pragmatic deixis in challenging tasks, with bilinguals exhibiting a higher sensitivity. The learning of adjectives is particularly challenging, but has rarely been investigated in bilingual children. In the present study we presented a pragmatic cue supporting the learning of novel adjectives to 32 Spanish–German bilingual and 28 German monolingual 5-year-olds. The children's responses to a descriptive hand gesture highlighting an object's property were measured behaviorally using a forced choice task and neurophysiologically through functional Near-Infrared Spectroscopy (fNIRS). While no group differences emerged on the behavioral level, fNIRS revealed a higher activation in bilingual than monolingual children in the vicinity of the posterior part of the right superior temporal sulcus (STS). This result supports the prominent role of the STS in processing pragmatic gestures and suggests heightened pragmatic sensitivity for bilingual children.
Methods can powerfully affect conclusions about infant experiences and learning. Data from naturalistic observations may paint a very different picture of learning and development from those based on structured tasks, as illustrated in studies of infant walking, object permanence, intention understanding, and so forth. Using language as a model system, we compared the speech of 40 mothers to their 13-month-old infants during structured play and naturalistic home routines. The contrasting methods yielded unique portrayals of infant language experiences, while simultaneously underscoring cross-situational correspondence at an individual level. Infants experienced substantially more total words and different words per minute during structured play than they did during naturalistic routines. Language input during structured play was consistently dense from minute to minute, whereas language during naturalistic routines showed striking fluctuations interspersed with silence. Despite these differences, infants' language experiences during structured play mirrored the peak language interactions infants experienced during naturalistic routines, and correlations between language inputs in the two conditions were strong. The implications of developmental methods for documenting the nature of experiences and individual differences are discussed.
Children learn the words of their native language(s) from interactions with their caregivers. Although previous research has found that the language children hear during those interactions predicts vocabulary outcomes, few studies have investigated how qualitative features of social interactions work together to affect children's vocabulary development, particularly for underresourced, language minoritized children. This study examined patterns of maternal interactive behaviors during toddlerhood in relation to children's later Spanish and English vocabulary development among 318 low-income, Mexican American families. Five maternal behaviors (acknowledging, elaborating, gaze, vocal appropriateness, and overriding) were coded from video recordings at age 24 months. At 36 and 54 months, child expressive vocabulary was assessed in both English and Spanish. Latent class analysis identified five distinct patterns of maternal interactive behaviors, which differentially supported or compromised child expressive language in English and Spanish. Implications of these findings for better understanding the role of caregiver interactions in dual language vocabulary development are discussed. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
Why do infants learn some words earlier than others? To explain how and when words are learned, existing theories of word learning prioritize visual information and draw mainly on lab-based studies of noun-to-object mapping. However, words that are more abstract than object nouns, such as uh-oh, hi, more, up, and all-gone, are typically among the first to appear in infants' vocabularies. We combined a behavioral experiment with naturalistic observational research to explore how infants learn and represent this understudied category of high-frequency, routine-based non-nouns, which we term ‘everyday words’. In Study 1, we found that conventional eye-tracking measures of comprehension were insufficient to capture 10- to 16-month-old infants' emerging understanding of everyday words. In Study 2, we analyzed the visual and social scenes surrounding caregivers' and infants' use of everyday words in a naturalistic video corpus. This ecologically-motivated research revealed that everyday words rarely co-occurred with consistent visual referents, making their early learnability difficult to reconcile with dominant word learning theories. Our findings instead point to complex patterns in the types of situations associated with everyday words that could contribute to their early representation in infants’ vocabularies. By leveraging both experimental and observational methods, this investigation underscores the value of using naturalistic data to broaden theories of early learning.
Humans primarily rely on language to communicate, on the basis of a shared understanding of the basic building blocks of communication: words. Do we mean the same things when we use the same words? Although cognitive neural research on semantics has revealed the common principles of word-meaning representation, the factors underlying the potential individual variations in word meanings are unknown. Here, we empirically characterized the intersubject consistency of 90 words across 20 adult subjects (10 female) using both behavioral measures (rating-based semantic-relationship patterns) and neuroimaging measures (word-evoked brain activity patterns). Across both the behavioral and neuroimaging experiments, we showed that the magnitude of individual disagreements on word meanings could be modeled on the basis of how much language or sensory experience is associated with a word and that this variation increases with word abstractness. Uncovering the cognitive and neural origins of word-meaning disagreements across individuals has implications for potential mechanisms to modulate such disagreements.
Many social and legal conflicts come down to differences in semantics. Yet, semantic variation between individuals and people’s awareness of this variation have been relatively neglected by experimental psychology. Here, across two experiments, we quantify the amount of agreement and disagreement between ordinary semantic concepts in thepopulation, as well as people’s meta-cognitive awareness of these differences. We collect similarity ratings and feature judgements, and analyze them using a non-parametricclustering scheme with an ecological statistical estimator to infer the number of different meanings for the same word that is present in the population. We find that typically atleast ten to twenty variants of meanings exist for even common nouns, but that people are unaware of this variation. Instead, people exhibit a strong bias to erroneously believe that other people share their particular semantics, pointing to one factor that likely interfereswith political and social discourse.
Objective:
Bilingual experiences are diverse, vibrant, and multidimensional. Yet, prior research has often homogenized bilingualism and based outcomes upon monolingual norms. Framing monolinguals as the norm distorts the reality of bilingual experiences. To promote a more diverse and inclusive study of bilingualism, we propose a theoretical and methodological paradigm shift. Bilinguals exist in different networks, cultural contexts, and individual and societal settings, all of which may lead to differential cognitive and linguistic outcomes that will be lost if left unexamined. Bilingual interactional contexts occur within extensive environmental and ecological systems, and may lead to different outcomes based on experiences within these systems. We seek to recognize these interactional contexts and how, as researchers, we can strive to better understand the complexities of bilingual populations.
Method:
We propose incorporating more diverse theoretical frameworks-including raciolinguistics, an intersectional resiliency perspective, and an ecological approach-so that researchers can begin to think about how bilingual experiences are shaped before study participants enter the lab. Included also are methodological considerations that will improve our understanding of bilinguals' intersectional experiences. We offer suggestions for becoming more diverse and inclusive in our research.
Conclusion:
We encourage scientists to take a more holistic and nuanced approach to understanding how individual and contextual factors affect our study populations. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
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. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
This longitudinal study documents the key role of early joint engagement in the language and early liter- acy development of Mexican American children from low-income households. This rapidly growing pop- ulation often faces challenges as sequential Spanish-English language learners. Videos of 121 mothers and their 2.5-year-old children interacting in Spanish for 15 min were recorded in 2009–2011 in the Dallas- Fort Worth metropolitan area. Researchers reliably rated general dyadic features of joint engagement— symbol-infused joint engagement, shared routines and rituals, and fluency and connectedness—that have been found to facilitate language development in young English-speaking children. The construct respeto, a valued aspect of traditional Latino parenting, was also rated using 2 culturally specific items—the par- ent’s calm authority and the child’s affiliative obedience. In addition, 3 individual contributions—maternal sensitivity, quality of maternal language input, and quality of child language production—were assessed. General features of joint engagement at 2.5 years predicted expressive and receptive language at 3.6 years and receptive language and early literacy at 7.3 years, accounting for unique variance over and above in- dividual contributions at 2.5 years, with some effects being stronger in girls than boys. The level of cul- turally specific joint engagement did not alter predictions made by general features of joint engagement. These findings highlight the importance of the quality of early communication for language and liter- acy success of Mexican American children from low-income households and demonstrate that culturally specific aspects of early interactions can align well with general features of joint engagement.
Infants use social cues like gaze and touch, typically during joint attention with caregivers, to decipher word meaning. Yet, studies on cues to word meaning primarily rely on structured tasks, where distractors are few compared to the countless objects that vie for attention in the home environment. Forty mothers and their 13‐month‐old infants were video‐recorded during home routines. From transcripts, 3,000+ “naming events” were coded for social (visual and manual) and contextual (location: room and place) cues to the object of reference. Mother and infant visual and tactile attention accompanied naming events with high regularity, and the objects mothers named were specific to room and place. Nested social and contextual cues infuse infants’ everyday language experiences in ways that help infants disambiguate word meaning.
This chapter presents the theory of syntactic bootstrapping. It shows fundamental problems with a theory of verb learning based solely on observations of the external world. It then shows how these problems can be overcome if those experiences are paired with information about the syntactic structure of the clause that the verb occurs in.
Children tend to produce words earlier when they are connected to a variety of other words along the phonological and semantic dimensions. Though these semantic and phonological connectivity effects have been extensively documented, little is known about their underlying developmental mechanism. One possibility is that learning is driven by lexical network growth where highly connected words in the child's early lexicon enable learning of similar words. Another possibility is that learning is driven by highly connected words in the external learning environment, instead of highly connected words in the early internal lexicon. The present study tests both scenarios systematically in both the phonological and semantic domains across 10 languages. We show that phonological and semantic connectivity in the learning environment drives growth in both production‐ and comprehension‐based vocabularies, even controlling for word frequency and length. This pattern of findings suggests a word learning process where children harness their statistical learning abilities to detect and learn highly connected words in the learning environment.
Early research has shown that there may be inconsistencies in how parents in different socioeconomic classes prepare their children to learn to read. Previous research has highlighted a “word gap” between children from low-income families and high-income families. Effects of this “word gap” are evident at the start of kindergarten and contribute to the current nationwide achievement gap in educational outcomes for low-income children. Providence Talks (PT) is a city-wide initiative launched in 2014 in an effort to close this gap. PT helps caretakers learn about the importance of speaking with their children at an early age and supports them in their ability to improve the language environments within their home. We uncover positive results for the efficacy of a citywide intervention dedicated to improving outcomes for vulnerable children across the urban landscape.
Many words are associated with more than a single meaning. Words are sometimes "ambiguous," applying to unrelated meanings, but the majority of frequent words are "polysemous" in that they apply to multiple related meanings. In a preregistered design that included 2 tasks, we tested adults' and 4.5- to 7-year-old children's ability to learn 4 novel polysemous words or 4 novel ambiguous words. Both children and adults demonstrated a polysemy over ambiguity learning advantage on each task after exposure, showing better learning of novel words with multiple related meanings than novel words with unrelated meanings. Stimuli in the polysemy condition were designed and then normed to guard against learners relying on a simple definition to distinguish the multiple target meanings for each word from foils. We retested available participants after a week-long delay without providing additional exposure and found that adults' performance remained strong in the polysemy condition in 1 task, and children's performance remained strong in the polysemy condition in both tasks. We conclude that participants are adept at learning polysemous words that vary along multiple dimensions. Current results are consistent with the idea that ambiguous meanings of a word compete, but polysemous meanings instead reinforce one another. (PsycINFO Database Record (c) 2020 APA, all rights reserved).
The cultural value of respeto (respect) is central to Latine parenting. Yet, how respeto manifests in the interactions of Latine parents and their young children remains unexamined. Low-income Mexican immigrant Spanish-speaking mothers and their 2.5-year-old toddlers (N = 128) were video-recorded during play (M
age
= 30.2 months, SD = 0.52), and two culturally informed items of respeto were coded: parent calm authority and child affiliative obedience. Respeto related to standard ratings of mother and child interactions (e.g., maternal sensitivity and child engagement) but also captured unique features of parent-child interactions. Respeto related to mothers' and toddlers' language production and discourse during the interaction, and explained unique variance in language variables above standard ratings of mother-child interaction. This is the first effort to document a culturally salient aspect of dyadic interaction in Mexican immigrant mothers and young children and to show that respeto relates to language use during mother- child interactions.
Can we predict the words a child is going to learn next given information about the words that a child knows now? Do different representations of a child’s vocabulary knowledge affect our ability to predict the acquisition of lexical items for individual children? Past research has often focused on population statistics of vocabulary growth rather than prediction of words an individual child is likely to learn next. We consider a neural network approach to predict vocabulary acquisition. Specifically, we investigate how best to represent the child’s current vocabulary in order to accurately predict future learning. The models we consider are based on qualitatively different sources of information: descriptive information about the child, the specific words a child knows, and representations that aim to capture the child’s aggregate lexical knowledge. Using longitudinal vocabulary data from children aged 15-36 months, we construct neural network models to predict which words are likely to be learned by a particular child in the coming month. Many models based on child-specific vocabulary information outperform models with child information only, suggesting that the words a child knows influence prediction of future language learning. These models provide an understanding of the role of current vocabulary knowledge on future lexical growth.
The LENA system has revolutionized research on language acquisition, providing both a wearable device to collect day-long recordings of children’s environments, and a set of automated outputs that process, identify, and classify speech using proprietary algorithms. This output includes information about input sources (e.g., adult male, electronics). While this system has been tested across a variety of settings, here we delve deeper into validating the accuracy and reliability of LENA’s automated diarization, i.e., tags of who is talking. Specifically, we compare LENA’s output with a gold standard set of manually generated talker tags from a dataset of 88 day-long recordings, taken from 44 infants at 6 and 7 months, which includes 57,983 utterances. We compare accuracy across a range of classifications from the original Lena Technical Report, alongside a set of analyses examining classification accuracy by utterance type (e.g., declarative, singing). Consistent with previous validations, we find overall high agreement between the human and LENA-generated speaker tags for adult speech in particular, with poorer performance identifying child, overlap, noise, and electronic speech (accuracy range across all measures: 0–92%). We discuss several clear benefits of using this automated system alongside potential caveats based on the error patterns we observe, concluding with implications for research using LENA-generated speaker tags.
What aspects of word meaning are important in early word learning and lexico-semantic network development? Adult lexico-semantic systems flexibly encode multiple types of semantic features, including functional, perceptual, taxonomic, and encyclopedic. However, various theoretical accounts of lexical development differ on whether and how these semantic properties of word meanings are initially encoded into young children's emerging lexico-semantic networks. Whereas some accounts highlight the importance of early perceptual versus conceptual properties, others posit that thematic or functional aspects of word meaning are primary relative to taxonomic knowledge. We seek to shed light on these debates with 2 modeling studies that explore patterns in early word learning using a large database of early vocabulary in 5,450 children, and a newly developed set of semantic features of early acquired nouns. In Study 1, we ask whether semantic properties of early acquired words relate to order in which these words are typically learned; Study 2 models normative lexico-semantic noun-feature network development compared to random network growth. Both studies provide converging evidence that perceptual properties of word meanings play a key role in early word learning and lexico-semantic network development. The findings lend support to theoretical accounts of language learning that highlight the importance of the child's perceptual experience.
There is considerable evidence that labeling supports infants’ object categorization. Yet in daily life, most of the category exemplars that infants encounter will remain unlabeled. Inspired by recent evidence from machine learning, we propose that infants successfully exploit this sparsely labeled input through “semi‐supervised learning.” Providing only a few labeled exemplars leads infants to initiate the process of categorization, after which they can integrate all subsequent exemplars, labeled or unlabeled, into their evolving category representations. Using a classic novelty preference task, we introduced 2‐year‐old infants (n = 96) to a novel object category, varying whether and when its exemplars were labeled. Infants were equally successful whether all exemplars were labeled (fully supervised condition) or only the first two exemplars were labeled (semi‐supervised condition), but they failed when no exemplars were labeled (unsupervised condition). Furthermore, the timing of the labeling mattered: when the labeled exemplars were provided at the end, rather than the beginning, of familiarization (reversed semi‐supervised condition), infants failed to learn the category. This provides the first evidence of semi‐supervised learning in infancy, revealing that infants excel at learning from exactly the kind of input that they typically receive in acquiring real‐world categories and their names.
This article is protected by copyright. All rights reserved.
It is well-known that children rapidly learn words, following a range of heuristics. What is less well appreciated is that—because most words are polysemous and have multiple meanings (e.g., “glass” can label a material and drinking vessel)—children will often be learning a new meaning for a known word, rather than an entirely new word. Across 4 experiments we show that children flexibly adapt a well-known heuristic—the shape bias—when learning polysemous words. Consistent with previous studies, we find that children and adults preferentially extend a new object label to other objects of the same shape. But we also find that when a new word for an object (“a gup”) has previously been used to label the material composing that object (“some gup”), children and adults override the shape bias, and are more likely to extend the object label by material (Experiments 1 and 3). Further, we find that, just as an older meaning of a polysemous word constrains interpretations of a new word meaning, encountering a new word meaning leads learners to update their interpretations of an older meaning (Experiment 2). Finally, we find that these effects only arise when learners can perceive that a word’s meanings are related, not when they are arbitrarily paired (Experiment 4). Together, these findings show that children can exploit cues from polysemy to infer how new word meanings should be extended, suggesting that polysemy may facilitate word learning and invite children to construe categories in new ways.
Learning the meanings of words involves not only linking individual words to referents but also building a network of connections among entities in the world, concepts, and words. Previous studies reveal that infants and adults track the statistical co-occurrence of labels and objects across multiple ambiguous training instances to learn words. However, it is less clear whether, given distributional or attentional cues, learners also encode associations among the novel objects. We investigated the consequences of two types of cues that highlighted object-object links in a cross-situational word learning task: distributional structure – how frequently the referents of novel words occurred together – and visual context – whether the referents were seen on matching backgrounds. Across three experiments, we found that in addition to learning novel words, adults formed connections between frequently co-occurring objects. These findings indicate that learners exploit statistical regularities to form multiple types of associations during word learning.
Semantic knowledge (or semantic memory) is knowledge we have about the world. For example, we know that knives are typically sharp, made of metal, and that they are tools used for cutting. To what kinds of experiences do we owe such knowledge? Most work has stressed the role of direct sensory and motor experiences. Another kind of experience, considerably less well understood, is our experience with language. We review two ways of thinking about the relationship between language and semantic knowledge: (i) language as mapping onto independently-acquired concepts, and (ii) language as a set of cues to meaning. We highlight some problems with the words-as-mappings view, and argue in favour of the words-as-cues alternative. We then review some surprising ways that language impacts semantic knowledge, and discuss how distributional semantics models can help us better understand its role. We argue that language has an abstracting effect on knowledge, helping to go beyond concrete experiences which are more characteristic of perception and action. We conclude by describing several promising directions for future research.
Researchers have focused for decades on how young children learn individual words. However, they have paid less attention to how children organize their word knowledge into the network of representations that underlies our ability to retrieve the right words efficiently and flexibly when we need them. Although methodological limits have made it difficult to study the development of lexical–semantic networks, in recent work with new paradigms, infants and toddlers have shown that they organize their vocabulary systematically via both categorical and associative relations from around age 2. Combined with work demonstrating that lexical–semantic relations affect early comprehension, production, and learning, this emerging area of research suggests that scientists and practitioners would benefit from more comprehensive theories of language learning that include not only vocabulary development, but also the development of lexical–semantic networks.
Objective:
To describe the trajectories of English and Spanish language growth in typically developing children from bilingual homes and compare those with the trajectories of English growth in children from monolingual homes, to assess effects of dual language exposure on language growth in typically developing children.
Study design:
Expressive vocabularies were assessed at 6-month intervals from age 30 to 60 months, in English for monolinguals and English and Spanish for bilinguals. Use of English and Spanish in the home was assessed via parental report.
Results:
Multilevel modeling, including parent education as a covariate, revealed that children from bilingual homes lagged 6 months to 1 year behind monolingual children in English vocabulary growth. The size of the lag was related to the relative amount of English use in the home, but the relation was not linear. Increments in English use conferred the greatest benefit most among homes with already high levels of English use. These homes also were likely to have 1 parent who was a native English speaker. Bilingual children showed stronger growth in English than in Spanish.
Conclusions:
Bilingual children can lag 6 months to 1 year behind monolingual children in normal English language development. Such lags may not necessarily signify clinically relevant delay if parents report that children also have skills in the home language. Shorter lags are associated with 2 correlated factors: more English exposure and more exposure from native English speakers. Early exposure to Spanish in the home does not guarantee acquisition of Spanish.
Objects in the world usually have names at different hierarchical levels (e.g., beagle, dog, animal). This research investigates adults' ability to use cross-situational statistics to simultaneously learn object labels at individual and category levels. The results revealed that adults were able to use co-occurrence information to learn hierarchical labels in contexts where the labels for individual objects and labels for categories were presented in completely separated blocks, in interleaved blocks, or mixed in the same trial. Temporal presentation schedules significantly affected the learning of individual object labels, but not the learning of category labels. Learners' subsequent generalization of category labels indicated sensitivity to the structure of statistical input.
Language exerts a powerful influence on our concepts. We review evidence documenting the developmental origins of a precocious link between language and object categories in very young infants. This collection of studies documents a cascading process in which early links between language and cognition provide the foundation for later, more precise ones. We propose that, early in life, language promotes categorization at least in part through its status as a social, communicative signal. But over the first year, infants home in on the referential power of language and, by their second year, begin teasing apart distinct kinds of names (e.g. nouns, adjectives) and their relation to distinct kinds of concepts (e.g. object categories, properties). To complement this proposal, we also relate this evidence to several alternative accounts of language's effect on categorization, appealing to similarity (‘labels-as-features’), familiarity (‘auditory overshadowing’), and communicative biases (‘natural pedagogy’).
Narratives in reminiscing contexts are an important tool in developing literacy abilities. We examined whether parents’ narrative participation styles in reminiscing contexts were related to gains in children's vocabulary and print knowledge. Participants were 210 low-income Chilean parents and their pre-kindergarten children (M = 53 months). Parents and children were videotaped discussing a past positive and negative experience at the beginning of pre-kindergarten. Children's vocabulary and print knowledge were assessed at the beginning and end of pre-kindergarten. Parents’ narrative styles in conversations about negative, but not positive, experiences had differential predictive power over children's print knowledge, but not vocabulary, longitudinally. Implications for policy makers, researchers and educators working with low-income Chilean families are discussed.
Infants' own activities create and actively select their learning experiences. Here we review recent models of embodied information seeking and curiosity-driven learning and show that these mechanisms have deep implications for development and evolution. We discuss how these mechanisms yield self-organized epigenesis with emergent ordered behavioral and cognitive developmental stages. We describe a robotic experiment that explored the hypothesis that progress in learning, in and for itself, generates intrinsic rewards: The robot learners probabilistically selected experiences according to their potential for reducing uncertainty. In these experiments, curiosity-driven learning led the robot learner to successively discover object affordances and vocal interaction with its peers. We explain how a learning curriculum adapted to the current constraints of the learning system automatically formed, constraining learning and shaping the developmental trajectory. The observed trajectories in the robot experiment share many properties with those in infant development, including a mixture of regularities and diversities in the developmental patterns. Finally, we argue that such emergent developmental structures can guide and constrain evolution, in particular with regard to the origins of language.
Curiosity is a basic element of our cognition, but its biological function, mechanisms, and neural underpinning remain poorly understood. It is nonetheless a motivator for learning, influential in decision-making, and crucial for healthy development. One factor limiting our understanding of it is the lack of a widely agreed upon delineation of what is and is not curiosity. Another factor is the dearth of standardized laboratory tasks that manipulate curiosity in the lab. Despite these barriers, recent years have seen a major growth of interest in both the neuroscience and psychology of curiosity. In this Perspective, we advocate for the importance of the field, provide a selective overview of its current state, and describe tasks that are used to study curiosity and information-seeking. We propose that, rather than worry about defining curiosity, it is more helpful to consider the motivations for information-seeking behavior and to study it in its ethological context.