summary: New research investigates how infants and young children acquire language. This study challenges preconceived notions about language development, especially in low-income households, by analyzing her one day’s worth of audio recordings of 1,001 children from a variety of backgrounds.
Research has shown that early comprehension begins around 6 to 7 months of age, with significant improvements in language comprehension occurring around the first birthday. The purpose of this study is to expand the scope of language development research to a more diverse population and to understand the mechanisms of language acquisition in children, including those who are deaf and visually impaired.
Important facts:
- Bergelson’s research refutes the assumption that socioeconomic status has a significant impact on children’s language development.
- Babies’ early language comprehension begins at six months of age and improves markedly by their first birthday.
- The study uses machine learning to analyze audio recordings of 1,001 children across 12 countries and 43 languages, providing a diverse and comprehensive dataset.
sauce: Harvard University
Growing up surrounded by Russian, Hebrew, and English, Erika Bergelson developed a passion for language development.
Her parents immigrated to Israel from the Soviet Union in the 1970s and started a family there. Bergelson and her youngest sibling were born in the 1980s after the family settled in Columbus, Ohio. Even back then, she noticed generational differences in grammar, accent, and vocabulary, and wondered how her children had outperformed the adults.
“Why does language acquisition actually make young minds, who are usually worse at everything, better at this particular process?” she recalled wondering.
Currently, the newly appointed associate professor of psychology studies how infants and young children learn language from the world around them. Developmental psychologists are particularly interested in parsing out the various theories that explain the beginning and eventual acquisition of language comprehension and production.
Bergelson’s most recent paper was published last month. Proceedings of the National Academy of Sciences, takes a global approach to developing and testing such theory, and the results refute common criticisms of low-income parents and caregivers.
“Our results certainly challenge some of the conventional wisdom in the U.S. policy field that families in certain socio-economic conditions provide more or less ‘good’ language input to their children. ” she said.
As a language scientist, Bergelson has a history of producing such myth-busting insights. Her first experiments on early word learning, conducted 15 years ago when she was a graduate student at the University of Pennsylvania, showed that her understanding begins at a much younger age than previously thought. It was revealed. “Around six or seven months, babies start to understand some really common nouns,” she said.
Scientists have long known that there is a rapid onset of language production around 18 months of age, Bergelson explained. In her follow-up study, she and her colleagues found similar qualitative improvements in language comprehension near children’s first birthdays, around the time their first real words arrive. . It’s like a child from about a year old barely understanding how language works to suddenly becoming a true communication partner.
Could this be because parents speak to older babies more or in different words? Bergelson studied this theory as a postdoctoral fellow and research professor at the University of Rochester, and has been researching this theory for ages 6 months to 18 months. He led the creation of a large-scale naturalistic dataset that tracks babies up to 10 months using audio and video recordings, eye tracking, and more.
“There does not seem to be any fundamental difference in how parents or caregivers interact with their 6-month-old and 12-month-old children,” she concluded.
With a grant from the National Institutes of Health, Bergelson’s new lab at Harvard University recently embarked on a project aimed at testing what she calls a “better learner model” of language acquisition. did. According to these theories, the turning point in understanding is thought to be due to the growth of the infant’s social, cognitive, or linguistic abilities, rather than simply the accumulation of input from the caregiver.
But what exactly are the skills that support word learning? Bergelson and colleagues plan to test indicators of comprehension that appear earlier than speaking itself, such as pointing or gazing in the direction of mentioned objects. is. This research has long-term potential to improve early intervention for children who struggle with language acquisition.
Bergelson also has a further goal of increasing the number of children that language scientists study. “The really important recent change in the field is a more serious consideration of the fact that we tend to study white middle-class Americans,” she says.
Her recent PNAS paper, co-authored with senior co-author Alejandrina Cristia of France’s École Normale Supérieure PSL, is based on a large sample of children aged 2 to 48 months. Throughout the day, her 1,001 children, representing 12 countries and 43 languages, recorded their babble and baby talk. Financial support for this research was provided by the National Science Foundation, the National Institutes of Health, and the National Endowment for the Humanities, among others.
Analysis of the recordings was completed with the help of machine learning. Bergelson called it a “coarse-grained” approach to studying the topic. “This is an algorithm’s estimate of how many words a child is hearing or saying,” she said. “But I think this is a complementary approach to work that is very time-consuming and has limited samples.”
The results showed that the main predictors of world language development are age, clinical factors such as prematurity and dyslexia, and how many words children receive from the world around them. . In contrast to previous studies, we found no effects related to gender, multilingualism, or socio-economics.
“In recent years, there has been a lot of discussion and debate in the literature about how socio-economic status is or is not related to language input and language output,” says the author. said Bergelson, who is personally fascinated by the babbling of babies in their early stages of development. Last year’s child.
“We looked at it in many, many, many different ways. Over this record of tens of thousands of hours of daily life, mothers who were more educated had children who said more words. No evidence of any kind was found that she gave birth to.”
Dr. Bergelson is also pursuing new research on language development in children who are deaf or visually impaired with a grant from the National Science Foundation. Blindness cases are particularly interesting, she noted.
“The language abilities of visually impaired adults are almost indistinguishable from those of sighted people,” she says. “But many of our theories about early language learning rely on children looking at others to refer to things in the world. So there’s the mystery: How does that happen? And what does that tell us about how language develops for everyone?”
About this neurodevelopment and language learning research news
author: Christy Desmith
sauce: Harvard University
contact: Christy DeSmith – Harvard University
image: Image credited to Neuroscience News
Original research: Closed access.
“Everyday language input and production of 1,001 children from six continents.Written by Erika Bergelson et al. PNAS
abstract
Everyday language input and production of 1,001 children from six continents.
Language is a universal human ability, easily acquired by young children who would otherwise struggle with many of the basics of survival. Still, language proficiency varies among individuals. Naturalistic and experimental observations suggest that children’s language abilities vary depending on factors such as socio-economic status and child gender. But which factors actually influence children’s everyday language use?
Here, we leverage voice technology in a big data approach to report on unique, cross-cultural, and diverse data sets. Over 2,500 d of child-centered audio recordings of 1,001 children aged 2 to 48 months from 12 countries. It spans six continents, across urban, farmer-gatherer, and subsistence farming contexts. As expected, age and language-related clinical risk and diagnosis predicted how much language (and language-like vocalizations) children produced.
Importantly, adult conversations were similar in children’s environments. Children who listened to more adult conversation produced more words. In contrast to previous conclusions based on more restricted sampling methods and a separate set of language proxies, socio-economic status (operationalized as maternal education) is significantly associated with child outcomes during the first 4 years of life. There was no association, nor was it associated with gender or multilingualism.
These findings from large-scale naturalistic data improve our understanding of which factors reliably predict variability in young learners’ speech behavior across a wide range of everyday contexts.