Modeling early phonetic acquisition from child-centered audio data.

Infants learn their native language(s) at an amazing speed. Before they even talk, their perception adapts to the language(s) they hear. However, the mechanisms responsible for this perceptual attunement and the circumstances in which it takes place remain unclear. This paper presents the first attempt to study perceptual attunement using ecological child-centered audio data. We show that a simple prediction algorithm exhibits perceptual attunement when applied on unrealistic clean audio-book data, but fails to do so when applied on ecologically-valid child-centered data. In the latter scenario, perceptual attunement only emerges when the prediction mechanism is supplemented with inductive biases that force the algorithm to focus exclusively on speech segments while learning speaker-, pitch-, and room-invariant representations. We argue these biases are plausible given previous research on infants and non-human animals. More generally, we show that what our model learns and how it develops through exposure to speech depends exquisitely on the details of the input signal. By doing so, we illustrate the importance of considering ecologically valid input data when modeling language acquisition.

Characterization of children's verbal input in a forager-farmer population using long-form audio recordings and diverse input definitions.

There is little systematically collected quantitative empirical data on how much linguistic input children in small-scale societies encounter, with some estimates suggesting low levels of directed speech. We report on an ecologically-valid analysis of speech experienced over the course of a day by young children (N = 24, 6-58 months old, 33% female) in a forager-horticulturalist population of lowland Bolivia. A permissive definition of input (i.e., including overlapping, background, and non-linguistic vocalizations) leads to massive changes in terms of input quantity, including a quadrupling of the estimate for overall input compared to a restrictive definition (only near and clear speech), while who talked to and around a focal child is relatively stable across input definitions. We discuss implications of these results for theoretical and empirical research into language acquisition.

Everyday language input and production in 1,001 children from six continents.

Language is a universal human ability, acquired readily by young children, who otherwise struggle with many basics of survival. And yet, language ability is variable across individuals. Naturalistic and experimental observations suggest that children's linguistic skills vary with factors like socioeconomic status and children's gender. But which factors really influence children's day-to-day language use? Here, we leverage speech technology in a big-data approach to report on a unique cross-cultural and diverse data set: >2,500 d-long, child-centered audio-recordings of 1,001 2- to 48-mo-olds from 12 countries spanning six continents across urban, farmer-forager, and subsistence-farming contexts. As expected, age and language-relevant clinical risks and diagnoses predicted how much speech (and speech-like vocalization) children produced. Critically, so too did adult talk in children's environments: Children who heard more talk from adults produced more speech. In contrast to previous conclusions based on more limited sampling methods and a different set of language proxies, socioeconomic status (operationalized as maternal education) was not significantly associated with children's productions over the first 4 y of life, and neither were gender or multilingualism. These findings from large-scale naturalistic data advance our understanding of which factors are robust predictors of variability in the speech behaviors of young learners in a wide range of everyday contexts.

Introducing Meta-analysis in the Evaluation of Computational Models of Infant Language Development.

Computational models of child language development can help us understand the cognitive underpinnings of the language learning process, which occurs along several linguistic levels at once (e.g., prosodic and phonological). However, in light of the replication crisis, modelers face the challenge of selecting representative and consolidated infant data. Thus, it is desirable to have evaluation methodologies that could account for robust empirical reference data, across multiple infant capabilities. Moreover, there is a need for practices that can compare developmental trajectories of infants to those of models as a function of language experience and development. The present study aims to take concrete steps to address these needs by introducing the concept of comparing models with large-scale cumulative empirical data from infants, as quantified by meta-analyses conducted across a large number of individual behavioral studies. We formalize the connection between measurable model and human behavior, and then present a conceptual framework for meta-analytic evaluation of computational models. We exemplify the meta-analytic model evaluation approach with two modeling experiments on infant-directed speech preference and native/non-native vowel discrimination.

Positive Valence Contributes to Hyperarticulation in Maternal Speech to Infants and Puppies.

Infant-directed speech often has hyperarticulated features, such as point vowels whose formants are further apart than in adult-directed speech. This increased "vowel space" may reflect the caretaker's effort to speak more clearly to infants, thus benefiting language processing. However, hyperarticulation may also result from more positive valence (e.g., speaking with positive vocal emotion) often found in mothers' speech to infants. This study was designed to replicate others who have found hyperarticulation in maternal speech to their 6-month-olds, but also to examine their speech to a non-human infant (i.e., a puppy). We rated both kinds of maternal speech for their emotional valence and recorded mothers' speech to a human adult. We found that mothers produced more positively valenced utterances and some hyperarticulation in both their infant- and puppy-directed speech, compared to their adult-directed speech. This finding promotes looking at maternal speech from a multi-faceted perspective that includes emotional state.

Diversity and representation in infant research: Barriers and bridges toward a globalized science of infant development.

Psychological researchers have become increasingly concerned with generalized accounts of human behavior based on narrow participant representation. This concern is particularly germane to infant research as findings from infant studies are often invoked to theorize broadly about the origins of human behavior. In this article, we examined participant diversity and representation in research published on infant development in four journals over the past decade. Sociodemographic data were coded for all articles reporting infant data published in Child Development, Developmental Science, Developmental Psychology, and Infancy between 2011 and 2022. Analyses of 1682 empirical articles, sampling approximately 1 million participants, revealed consistent under-reporting of sociodemographic information. For studies that reported sociodemographic characteristics, there was an unwavering skew toward White infants from North America/Western Europe. To address a lack of diversity in infant studies and its scientific impact, a set of principles and practices are proposed to advance toward a more globally representative science.

Combining observational and experimental approaches to the development of language and communication in rural samples: Opportunities and challenges.

Multiple approaches - including observational and experimental - are necessary to articulate powerful theories of learning. Our field's key questions, which rely on these varied methods, are still open. How do children perceive and produce language? What do they encounter in their linguistic input? What does the learner bring to the task of acquisition? Considerable progress has been made for the development of spoken English (especially by North American learners). Yet there is still a great deal to discover about how children in other populations proceed, especially populations in rural settings. To examine language learning in these populations, we need a multi-method approach. However, adapting and integrating methods, particularly experimental ones, to new settings can present immense challenges. In this paper, we discuss the opportunities and challenges facing researchers who aim to use a multimethodological approach in rural samples, and what the field of language acquisition can do to promote such work.

Vocal input and output among infants in a multilingual context: Evidence from long-form recordings in Vanuatu.

What are the vocal experiences of children growing up on Malakula island, Vanuatu, where multilingualism is the norm? Long-form audio-recordings captured spontaneous speech behavior by, and around, 38 children (5-33 months, 23 girls) from 11 villages. Automated analyses revealed most children's vocal input came from female adults and other children's voices, with small contributions from male adult voices. The greatest changes with age involved an increase in the input vocalizations from other children. Total input (collapsing across child-directed and overheard speech, and across languages) was ∼11 min per hour, which was at least 5 min (31%) lower than that found in other populations studied using comparable methods in previous literature, as well as in archival American data analyzed with the same algorithm. In contrast, children's own vocalization counts were two to four times higher than previous reports for North-American English-learning monolingual infants at matched ages, and comparable to estimates from archival American data, consistent with a resilient language-learning cognitive system for this aspect of vocal development. The strongest association between input and output was with vocalizations by other children, rather than those by adults, which is consistent with research in anthropology but less so with current theoretical trends in developmental psychology. These results invite further research in populations that are under-represented in developmental science. RESEARCH HIGHLIGHTS: Combining long-form recordings with automated analyses, we estimated infants potentially exposed to ∼2.6 languages heard ∼11 min of speech per hour. Infants' input was dominated by vocalizations from female adults and from other children, particularly for the oldest infants in our sample. The strongest association between children's own vocalization counts and input counts was with those of other children, and not those with adults. Results invite further research on individual, group, and population variability in input quantity and composition, and its potential effects on vocal development.

Infants' preference for speech is stable across the first year of life: Meta-analytic evidence.

Previous work suggested that humans' sophisticated speech perception abilities stem from an early capacity to pay attention to speech in the auditory environment. What are the roots of this early preference? We assess the extent to which it is due to it being a vocal sound, a natural sound, and a familiar sound through a meta-analytic approach, classifying experiments as a function of whether they used native or foreign speech and whether the competitor, against which preference is tested, was vocal or non-vocal, natural or artificial. We also tested for the effect of age. Synthesizing data from 791 infants across 39 experiments, we found a medium effect size, confirming at the scale of the literature that infants reliably prefer speech over other sounds. This preference was not significantly moderated by the language used, vocal quality, or naturalness of the competitor, nor by infant age. The current body of evidence appears most compatible with the hypothesis that speech is preferred consistently as such and not just due to its vocal, natural, or familiar nature. We discuss limitations of the extant body of work on speech preference, including evidence consistent with a publication bias and low representation of certain stimuli types and ages.

A systematic review suggests marked differences in the prevalence of infant-directed vocalization across groups of populations.

Anthropological reports have long suggested that speaking to young children is very infrequent in certain populations (notably farming ones), which is in line with scattered quantitative studies. A systematic review was undertaken to use available literature in order to estimate the extent of population variation. Database searches, expert lists, and citation searches led to the discovery of 29 reports on the frequency of vocalizations directed to infants aged 24 months or younger, based on systematic observations of spontaneous activity in the infant's natural environment lasting at least 30 min in length. Together, these studies provide evidence on 1314 infants growing up in a range of communities (urban, foraging, farming). For populations located outside of North America, the frequency with which vocalization was directed to urban infants was much higher than that for rural infants (including both foraging and farming, medians = 12.6 vs. 3.6% of observations contained infant-directed vocalization behaviors). We benchmarked this effect against socio-economic status (SES) variation in the United States, which was much smaller. Infants in high SES American homes were spoken to only slightly more frequently than those in low SES homes (medians = 16.4 vs. 15.1% of observations contained infant-directed vocalization behaviors). Although published research represents a biased sample of the world's populations, these results invite further cross-population research to understand the causes and effects of such considerable population group differences.

How Effective Altruism Can Help Psychologists Maximize Their Impact.

Although many psychologists are interested in making the world a better place through their work, they are often unable to have the impact that they would like. Here, we suggest that both individuals and psychology as a field can better improve human welfare by incorporating ideas from effective altruism, a growing movement whose members aim to do the most good by using science and reason to inform their efforts. In this article, we first briefly introduce effective altruism and review important principles that can be applied to how psychologists approach their work, such as the importance, tractability, and neglectedness framework. We then review how effective altruism can inform individual psychologists' choices. Finally, we close with a discussion of ideas for how psychology, as a field, can increase its positive impact. By applying insights from effective altruism to psychological science, we aim to integrate a new theoretical framework into psychological science, stimulate new areas of research, start a discussion on how psychology can maximize its impact, and inspire the psychology community to do the most good.

Correction: Infant-directed input and literacy effects on phonological processing: Non-word repetition scores among the Tsimane'.

[This corrects the article DOI: 10.1371/journal.pone.0237702.].

Factors structuring lexical development in toddlers: The effects of parental education, language exposure, and age.

A growing body of research suggests that individual variation in young children's word comprehension (indexed by response times and accuracy) is structured and meaningful. In this paper, we assess how children's word comprehension correlates with three factors: socio-economic status (indexed by maternal education), lingual status (based on language exposure), and age. We present results from 91 2- to 3-year-old children using a paired forced-choice task built on a child-friendly touch screen. Effects associated with maternal education and exposure to the tested language (French) were small, and they were greater for accuracy than response times. This pattern of results is compatible with an interpretation whereby the greatest effects of these two variables are on cumulative knowledge (vocabulary size) rather than on processing. Effects for age were larger and affected both accuracy and response times. Finally, response time variation did not mediate the effects of socio-economic status on accuracy or vice versa.

Using big data from long-form recordings to study development and optimize societal impact.

Big data are everywhere. In this chapter, we focus on one source: long-form, child-centered recordings collected using wearable technologies. Because these recordings are simultaneously unobtrusive and encompassing, they may be a breakthrough technology for clinicians and researchers from several diverse fields. We demonstrate this possibility by outlining three applications for the recordings-clinical treatment, large-scale interventions, and language documentation-where we see the greatest potential. We argue that incorporating these recordings into basic and applied research will result in more equitable treatment of patients, more reliable measurements of the effects of interventions on real-world behavior, and deeper scientific insights with less observational bias. We conclude by outlining a proposal for a semistructured online platform where vast numbers of long-form recordings could be hosted and more representative, less biased algorithms could be trained.

Socioeconomic status correlates with measures of Language Environment Analysis (LENA) system: a meta-analysis.

Using a meta-analytic approach, we evaluate the association between socioeconomic status (SES) and children's experiences measured with the Language Environment Analysis (LENA) system. Our final analysis included 22 independent samples, representing data from 1583 children. A model controlling for LENATM measures, age and publication type revealed an effect size of r z = .186, indicating a small effect of SES on children's language experiences. The type of LENA metric measured emerged as a significant moderator, indicating stronger effects for adult word counts than child vocalization counts. These results provide important evidence for the strength of association between SES and children's everyday language experiences as measured with an unobtrusive recording analyzed automatically in a standardized fashion.

Does morphological complexity affect word segmentation? Evidence from computational modeling.

How can infants detect where words or morphemes start and end in the continuous stream of speech? Previous computational studies have investigated this question mainly for English, where morpheme and word boundaries are often isomorphic. Yet in many languages, words are often multimorphemic, such that word and morpheme boundaries do not align. Our study employed corpora of two languages that differ in the complexity of inflectional morphology, Chintang (Sino-Tibetan) and Japanese (in Experiment 1), as well as corpora of artificial languages ranging in morphological complexity, as measured by the ratio and distribution of morphemes per word (in Experiments 2 and 3). We used two baselines and three conceptually diverse word segmentation algorithms, two of which rely purely on sublexical information using distributional cues, and one that builds a lexicon. The algorithms' performance was evaluated on both word- and morpheme-level representations of the corpora. Segmentation results were better for the morphologically simpler languages than for the morphologically more complex languages, in line with the hypothesis that languages with greater inflectional complexity could be more difficult to segment into words. We further show that the effect of morphological complexity is relatively small, compared to that of algorithm and evaluation level. We therefore recommend that infant researchers look for signatures of the different segmentation algorithms and strategies, before looking for differences in infant segmentation landmarks across languages varying in complexity.

How much does prosody help word segmentation? A simulation study on infant-directed speech.

Infants come to learn several hundreds of word forms by two years of age, and it is possible this involves carving these forms out from continuous speech. It has been proposed that the task is facilitated by the presence of prosodic boundaries. We revisit this claim by running computational models of word segmentation, with and without prosodic information, on a corpus of infant-directed speech. We use five cognitively-based algorithms, which vary in whether they employ a sub-lexical or a lexical segmentation strategy and whether they are simple heuristics or embody an ideal learner. Results show that providing expert-annotated prosodic breaks does not uniformly help all segmentation models. The sub-lexical algorithms, which perform more poorly, benefit most, while the lexical ones show a very small gain. Moreover, when prosodic information is derived automatically from the acoustic cues infants are known to be sensitive to, errors in the detection of the boundaries lead to smaller positive effects, and even negative ones for some algorithms. This shows that even though infants could potentially use prosodic breaks, it does not necessarily follow that they should incorporate prosody into their segmentation strategies, when confronted with realistic signals.

Child-directed and overheard input from different speakers in two distinct cultures.

Despite the fact that in most communities interaction occurs between the child and multiple speakers, most previous research on input to children focused on input from mothers. We annotated recordings of Sesotho-learning toddlers living in non-industrial Lesotho in South Africa, and French-learning toddlers living in urban regions in France. We examined who produced the input (mothers, other children, adults), how much input was child directed, and whether and how it varied across speakers. As expected, mothers contributed most of the input in the French recordings. However, in the Sesotho recordings, input from other children was more common than input from mothers or other adults. Child-directed speech from all speakers in both cultural groups showed similar qualitative modifications. Our findings suggest that input from other children is prevalent and has similar features as child-directed from adults described in previous work, inviting cross-cultural research into the effects of input from other children.

SCALa: A blueprint for computational models of language acquisition in social context.

Theories and data on language acquisition suggest a range of cues are used, ranging from information on structure found in the linguistic signal itself, to information gleaned from the environmental context or through social interaction. We propose a blueprint for computational models of the early language learner (SCALa, for Socio-Computational Architecture of Language Acquisition) that makes explicit the connection between the kinds of information available to the social learner and the computational mechanisms required to extract language-relevant information and learn from it. SCALa integrates a range of views on language acquisition, further allowing us to make precise recommendations for future large-scale empirical research.

Describing Vocalizations in Young Children: A Big Data Approach Through Citizen Science Annotation.

Purpose Recording young children's vocalizations through wearables is a promising method to assess language development. However, accurately and rapidly annotating these files remains challenging. Online crowdsourcing with the collaboration of citizen scientists could be a feasible solution. In this article, we assess the extent to which citizen scientists' annotations align with those gathered in the lab for recordings collected from young children. Method Segments identified by Language ENvironment Analysis as produced by the key child were extracted from one daylong recording for each of 20 participants: 10 low-risk control children and 10 children diagnosed with Angelman syndrome, a neurogenetic syndrome characterized by severe language impairments. Speech samples were annotated by trained annotators in the laboratory as well as by citizen scientists on Zooniverse. All annotators assigned one of five labels to each sample: Canonical, Noncanonical, Crying, Laughing, and Junk. This allowed the derivation of two child-level vocalization metrics: the Linguistic Proportion and the Canonical Proportion. Results At the segment level, Zooniverse classifications had moderate precision and recall. More importantly, the Linguistic Proportion and the Canonical Proportion derived from Zooniverse annotations were highly correlated with those derived from laboratory annotations. Conclusions Annotations obtained through a citizen science platform can help us overcome challenges posed by the process of annotating daylong speech recordings. Particularly when used in composites or derived metrics, such annotations can be used to investigate early markers of language delays.