Publications

Abstract

While there are well-known demonstrations that children can use distributional information to acquire multiple components of language, the underpinnings of these achievements are unclear. In the current paper, we investigate the potential pre-requisites for a distributional learning model that can explain how children learn their first words. We review existing literature and then present the results of a series of computational simulations with Vector Space Models, a type of distributional semantic model used in Computational Linguistics, which we evaluate against vocabulary acquisition data from children. We focus on nouns and verbs, and we find that: (i) a model with flexibility to adjust for the frequency of events provides a better fit to the human data, (ii) the influence of context words is very local, especially for nouns, and (iii) words that share more contexts with other words are harder to learn.

Abstract

Language is a universal human ability, acquired readily by young children, whootherwise struggle with many basics of survival. And yet, language ability is variableacross individuals. Naturalistic and experimental observations suggest that children’slinguistic skills vary with factors like socioeconomic status and children’s gender.But which factors really influence children’s day-to-day language use? Here, weleverage speech technology in a big-data approach to report on a unique cross-culturaland diverse data set: >2,500 d-long, child-centered audio-recordings of 1,001 2- to48-mo-olds from 12 countries spanning six continents across urban, farmer-forager,and subsistence-farming contexts. As expected, age and language-relevant clinical risksand diagnoses predicted how much speech (and speech-like vocalization) childrenproduced. Critically, so too did adult talk in children’s environments: Children whoheard more talk from adults produced more speech. In contrast to previous conclusionsbased on more limited sampling methods and a different set of language proxies,socioeconomic status (operationalized as maternal education) was not significantlyassociated with children’s productions over the first 4 y of life, and neither weregender or multilingualism. These findings from large-scale naturalistic data advanceour understanding of which factors are robust predictors of variability in the speechbehaviors of young learners in a wide range of everyday contexts

Abstract

Eye gaze is a powerful ostensive cue in infant-caregiver interactions, with demonstrable effects on language acquisition. While the link between gaze following and later vocabulary is well-established, the effects of eye gaze on other aspects of language, such as speech processing, are less clear. In this EEG study, we examined the effects of the speaker’s eye gaze on ten-month-old infants’ neural tracking of naturalistic audiovisual speech, a marker for successful speech processing. Infants watched videos of a speaker telling stories, addressing the infant with direct or averted eye gaze. We assessed infants’ speech-brain coherence at stress (1–1.75 Hz) and syllable (2.5–3.5 Hz) rates, tested for differences in attention by comparing looking times and EEG theta power in the two conditions, and investigated whether neural tracking predicts later vocabulary. Our results showed that infants’ brains tracked the speech rhythm both at the stress and syllable rates, and that infants’ neural tracking at the syllable rate predicted later vocabulary. However, speech-brain coherence did not significantly differ between direct and averted gaze conditions and infants did not show greater attention to direct gaze. Overall, our results suggest significant neural tracking at ten months, related to vocabulary development, but not modulated by speaker’s gaze.

Abstract

At the group level, children exposed to certain health and demographic risk factors, and who have delayed language in early childhood are, more likely to have language problems later in childhood. However, it is unclear whether we can use these risk factors to predict whether an individual child is likely to develop problems with language (e.g., be diagnosed with a developmental language disorder). We tested this in a sample of 146 children who took part in the UK-CDI norming project. When the children were 15–18 months old, 1,210 British parents completed: (a) the UK-CDI (a detailed assessment of vocabulary and gesture use) and (b) the Family Questionnaire (questions about health and demographic risk factors). When the children were between 4 and 6  years, 146 of the same parents completed a short questionnaire that assessed (a) whether children had been diagnosed with a disability that was likely to affect language proficiency (e.g., developmental disability, language disorder, hearing impairment), but (b) also yielded a broader measure: whether the child’s language had raised any concern, either by a parent or professional. Discriminant function analyses were used to assess whether we could use different combinations of 10 risk factors, together with early vocabulary and gesture scores, to identify children (a) who had developed a language-related disability by the age of 4–6 years (20 children, 13.70% of the sample) or (b) for whom concern about language had been expressed (49 children; 33.56%). The overall accuracy of the models, and the specificity scores were high, indicating that the measures correctly identified those children without a language-related disability and whose language was not of concern. However, sensitivity scores were low, indicating that the models could not identify those children who were diagnosed with a language-related disability or whose language was of concern. Several exploratory analyses were carried out to analyse these results further. Overall, the results suggest that it is difficult to use parent reports of early risk factors and language in the first 2 years of life to predict which children are likely to be diagnosed with a language-related disability. Possible reasons for this are discussed.

Abstract

The present study investigates the developmental trajectory of abstract representations for syntactic structures in children. In a structural priming experiment on the dative alternation in German, we primed children from three different age groups (3–4 years, 5–6 years, 7–8 years) and adults with double object datives (Dora sent Boots the rabbit) or prepositional object datives (Dora sent the rabbit to Boots). Importantly, the prepositional object structure in German is dispreferred and only rarely encountered by young children. While immediate as well as cumulative structural priming effects occurred across all age groups, these effects were strongest in the 3- to 4-year-old group and gradually decreased with increasing age. These results suggest that representations in young children are less stable than in adults and, therefore, more susceptible to adaptation both immediately and across time, presumably due to stronger surprisal. Lexical boost effects, in contrast, were not present in 3- to 4-year-olds but gradually emerged with increasing age, possibly due to limited working-memory capacity in the younger child groups.

Abstract

There is a wealth of evidence demonstrating that executive function (EF) abilities are positively associated with language development during the preschool years, such that children with good executive functions also have larger vocabularies. However, why this is the case remains to be discovered. In this study, we focused on the hypothesis that sentence processing abilities mediate the association between EF skills and receptive vocabulary knowledge, in that the speed of language acquisition is at least partially dependent on a child’s processing ability, which is itself dependent on executive control. We tested this hypothesis in longitudinal data from a cohort of 3- and 4-year-old children at three age points (37, 43, and 49 months). We found evidence, consistent with previous research, for a significant association between three EF skills (cognitive flexibility, working memory [as measured by the Backward Digit Span], and inhibition) and receptive vocabulary knowledge across this age range. However, only one of the tested sentence processing abilities (the ability to maintain multiple possible referents in mind) significantly mediated this relationship and only for one of the tested EFs (inhibition). The results suggest that children who are better able to inhibit incorrect responses are also better able to maintain multiple possible referents in mind while a sentence unfolds, a sophisticated sentence processing ability that may facilitate vocabulary learning from complex input.

Abstract

For shared book reading to be effective for language development, the adult and child need to be highly engaged. The current paper adopted a mixed-methods approach to investigate caregiver’s language-boosting behaviours and children’s engagement during shared book reading. The results revealed there were more instances of joint attention and caregiver’s use of prompts during moments of higher engagement. However, instances of most language-boosting behaviours were similar across episodes of higher and lower engagement. Qualitative analysis assessing the link between children’s engagement and caregiver’s use of speech acts, revealed that speech acts do seem to contribute to high engagement, in combination with other aspects of the interaction.

Abstract

We investigated whether learning an artificial language at 17 months was predictive of children’s natural language vocabulary and grammar skills at 54 months. Children at 17 months listened to an artificial language containing non-adjacent dependencies, and were then tested on their learning to segment and to generalise the structure of the language. At 54 months, children were then tested on a range of standardised natural language tasks that assessed receptive and expressive vocabulary and grammar. A structural equation model demonstrated that learning the artificial language generalisation at 17 months predicted language abilities – a composite of vocabulary and grammar skills – at 54 months, whereas artificial language segmentation at 17 months did not predict language abilities at this age. Artificial language learning tasks – especially those that probe grammar learning – provide a valuable tool for uncovering the mechanisms driving children’s early language development.

Abstract

Caregiver abstract talk during shared reading predicts preschool-age children’s vocabulary development. However, previous research has focused on level of abstraction with less consideration of the style of extratextual talk. Here, we investigated the relation between these two dimensions of extratextual talk, and their contributions to variance in children’s vocabulary skills. Caregiver level of abstraction was associated with an interactive reading style. Controlling for socioeconomic status and child age, high interactivity predicted children’s concurrent vocabulary skills whereas abstraction did not. Controlling for earlier vocabulary skills, neither dimension of the extratextual talk predicted later vocabulary. Theoretical and practical relevance are discussed.

Abstract

Children’s ability to share attention with another social partner (i.e., joint attention) has been found to support language development. Despite the large amount of research examining the effects of joint attention on language in hearing population, little is known about how deaf children learning sign languages achieve joint attention with their caregivers during natural social interaction and how caregivers provide and scaffold learning opportunities for their children. The present study investigates the properties and timing of joint attention surrounding familiar and novel naming events and their relationship to children’s vocabulary. Naturalistic play sessions of caretaker-child-dyads using American Sign Language were analyzed in regards to naming events of either familiar or novel object labeling events and the surrounding joint attention events. We observed that most naming events took place in the context of a successful joint attention event and that sign familiarity was related to the timing of naming events within the joint attention events. Our results suggest that caregivers are highly sensitive to their child’s visual attention in interactions and modulate joint attention differently in the context of naming events of familiar vs. novel object labels.

Abstract

We used a multi-method approach to investigate how children avoid (or retreat from) argument structure overgeneralisation errors (e.g., *You giggled me). Experiment 1 investigated how semantic and statistical constraints (preemption and entrenchment) influence children’s and adults’ judgments of the grammatical acceptability of 120 verbs in transitive and intransitive sentences. Experiment 2 used syntactic priming to elicit overgeneralisation errors from children (aged 5–6) to investigate whether the same constraints operate in production. For judgments, the data showed effects of preemption, entrenchment, and semantics for all ages. For production, only an effect of preemption was observed, and only for transitivisation errors with intransitive-only verbs (e.g., *The man laughed the girl). We conclude that preemption, entrenchment, and semantic effects are real, but are obscured by particular features of the present production task.

Abstract

Eye gaze is a ubiquitous cue in child-caregiver interactions and infants are highly attentive to eye gaze from very early on. However, the question of why infants show gaze-sensitive behavior, and what role this sensitivity to gaze plays in their language development, is not yet well-understood. To gain a better understanding of the role of eye gaze in infants’ language learning, we conducted a broad systematic review of the developmental literature for all studies that investigate the role of eye gaze in infants’ language development. Across 77 peer-reviewed articles containing data from typically-developing human infants (0-24 months) in the domain of language development we identified two broad themes. The first tracked the effect of eye gaze on four developmental domains: (1) vocabulary development, (2) word-object mapping, (3) object processing, and (4) speech processing. Overall, there is considerable evidence that infants learn more about objects and are more likely to form word-object mappings in the presence of eye gaze cues, both of which are necessary for learning words. In addition, there is good evidence for longitudinal relationships between infants’ gaze following abilities and later receptive and expressive vocabulary. However, many domains (e.g. speech processing) are understudied; further work is needed to decide whether gaze effects are specific to tasks such as word-object mapping, or whether they reflect a general learning enhancement mechanism. The second theme explored the reasons why eye gaze might be facilitative for learning, addressing the question of whether eye gaze is treated by infants as a specialized socio-cognitive cue. We concluded that the balance of evidence supports the idea that eye gaze facilitates infants’ learning by enhancing their arousal, memory and attentional capacities to a greater extent than other low-level attentional cues. However, as yet, there are too few studies that directly compare the effect of eye gaze cues and non-social, attentional cues for strong conclusions to be drawn. We also suggest there might be a developmental effect, with eye gaze, over the course of the first two years of life, developing into a truly ostensive cue that enhances language learning across the board.

Abstract

In the previous decade, dozens of studies involving thousands of children across several research disciplines have made use of a combined daylong audio-recorder and automated algorithmic analysis called the LENAⓇ system, which aims to assess children’s language environment. While the system’s prevalence in the language acquisition domain is steadily growing, there are only scattered validation efforts on only some of its key characteristics. Here, we assess the LENAⓇ system’s accuracy across all of its key measures: speaker classification, Child Vocalization Counts (CVC), Conversational Turn Counts (CTC), and Adult Word Counts (AWC). Our assessment is based on manual annotation of clips that have been randomly or periodically sampled out of daylong recordings, collected from (a) populations similar to the system’s original training data (North American English-learning children aged 3-36 months), (b) children learning another dialect of English (UK), and (c) slightly older children growing up in a different linguistic and socio-cultural setting (Tsimane’ learners in rural Bolivia). We find reasonably high accuracy in some measures (AWC, CVC), with more problematic levels of performance in others (CTC, precision of male adults and other children). Statistical analyses do not support the view that performance is worse for children who are dissimilar from the LENAⓇ original training set. Whether LENAⓇ results are accurate enough for a given research, educational, or clinical application depends largely on the specifics at hand. We therefore conclude with a set of recommendations to help researchers make this determination for their goals.

Abstract

Understanding the mechanisms that drive variation in children’s language acquisition requires large, population-representative datasets of children’s word learning across development. Parent report measures such as the MacArthur-Bates Communicative Development Inventories (CDI) are commonly used to collect such data, but the traditional paper-based forms make the curation of large datasets logistically challenging. Many CDI datasets are thus gathered using convenience samples, often recruited from communities in proximity to major research institutions. Here, we introduce Web-CDI, a web-based tool which allows researchers to collect CDI data online. Web-CDI contains functionality to collect and manage longitudinal data, share links to test administrations, and download vocabulary scores. To date, over 3,500 valid Web-CDI administrations have been completed. General trends found in past norming studies of the CDI are present in data collected from Web-CDI: scores of children’s productive vocabulary grow with age, female children show a slightly faster rate of vocabulary growth, and participants with higher levels of educational attainment report slightly higher vocabulary production scores than those with lower levels of education attainment. We also report results from an effort to oversample non-white, lower-education participants via online recruitment (N = 241). These data showed similar demographic trends to the full sample but this effort resulted in a high exclusion rate. We conclude by discussing implications and challenges for the collection of large, population-representative datasets.

Abstract

By the end of their first year, infants can interpret many different types of complex dynamic visual events, such as caused-motion, chasing, and goal-directed action. Infants of this age are also in the early stages of vocabulary development, producing their first words at around 12 months. The present work examined whether there are meaningful individual differences in infants’ ability to represent dynamic causal events in visual scenes, and whether these differences influence vocabulary development. As part of the longitudinal Language 0–5 Project, 78 10-month-old infants were tested on their ability to interpret three dynamic motion events, involving (a) caused-motion, (b) chasing behaviour, and (c) goal-directed movement. Planned analyses found that infants showed evidence of understanding the first two event types, but not the third. Looking behaviour in each task was not meaningfully related to vocabulary development, nor were there any correlations between the tasks. The results of additional exploratory analyses and simulations suggested that the infants’ understanding of each event may not be predictive of their vocabulary development, and that looking times in these tasks may not be reliably capturing any meaningful individual differences in their knowledge. This raises questions about how to convert experimental group designs to individual differences measures, and how to interpret infant looking time behaviour.

Abstract

A key omission from many accounts of children’s early word learning is the linguistic knowledge that the child has acquired up to the point when learning occurs. We simulate this knowledge using a computational model that learns phoneme and word sequence knowledge from naturalistic language corpora. We show how this simple model is able to account for effects of word length, word frequency, neighborhood density and phonotactic probability on children’s early word learning. Moreover, we show how effects of neighborhood density and phonotactic probability on word learning are largely influenced by word length, with our model being able to capture all effects. We then use predictions from the model to show how the ease by which a child learns a new word from maternal input is directly influenced by the phonological knowledge that the child has acquired from other words up to the point of encountering the new word. There are major implications of this work: models and theories of early word learning need to incorporate existing sublexical and lexical knowledge in explaining developmental change while well-established indices of word learning are rejected in favor of phonological knowledge of varying grain sizes.

Abstract

The purpose of this project was to investigate the effectiveness of a language-focused professional development programme on the knowledge and behaviour of preschool practitioners (sometimes called early years practitioners) in the UK. In Study 1 we determined whether the training received by practitioners is effective in improving their knowledge of how to support children’s language and communicative development. In Study 2 we tested whether trained practitioners, and practitioners from centres with embedded Language Champions, were able to implement the techniques they had been taught. For this, we video-recorded practitioners interacting, one to one, with 2- and 3–4-year-old children in their centres. We conclude that (1) practitioners retain the knowledge they have been taught, both about how children learn and about how to promote this learning, and that (2), in some respects, this knowledge translates well into practice; practitioners in centres with embedded Language Champions and trained practitioners used language-enriching behaviours when interacting with children more often than did untrained practitioners. We discuss how the translation of some techniques into overt behaviour could be made more effective

Abstract

Language users encounter words in at least two different modalities. Arguably, the most frequent encounters are in spoken or written form. Previous research has shown that – compared to the spoken modality – written language features more difficult words. Thus, frequent reading might have effects on word recognition. In the present study, we investigated 1) whether input modality (spoken, written, or bimodal) has an effect on word recognition accuracy, 2) whether this modality effect interacts with word difficulty, 3) whether the interaction of word difficulty and reading experience impacts word recognition accuracy, and 4) whether this interaction is influenced by input modality. To do so, we re-analysed a dataset that was collected in the context of a vocabulary test development to assess in which modality test words should be presented. Participants had carried out a word recognition task, where non-words and words of varying difficulty were presented in auditory, visual and audio-visual modalities. In addition to this main experiment, participants had completed a receptive vocabulary and an author recognition test to measure their reading experience. Our re-analyses did not reveal evidence for an effect of input modality on word recognition accuracy, nor for interactions with word difficulty or language experience. Word difficulty interacted with reading experience in that frequent readers were more accurate in recognizing difficult words than individuals who read less frequently. Practical implications are discussed.