A new study reveals an often-overlooked process in language learning: children frequently begin by grasping whole phrases and only later identify the individual components within them. This “whole-to-part” approach suggests that children have an inherent bias toward holistic learning, challenging the traditional view that language is usually acquired by assembling smaller parts into larger structures. The findings shed light on the flexibility and resilience of language acquisition across diverse contexts, including situations where formal language input is limited.
A recent study led by Professor Inbal Arnon from the Department of Psychology at the Hebrew University of Jerusalem, in collaboration with Professor Susan Goldin-Meadow from the Department of Psychology at the University of Chicago, has unveiled new insight on how children learn language. Traditionally, it has been thought that children approach language by building larger structures, such as words and sentences, out of smaller units, like syllables. Arnon’s past work however, highlights an equally vital process: children sometimes start with whole phrases and later identify the parts within them. In this paper, the researchers show this process in children not exposed to linguistic input: this challenges the idea that children’s approach to language is strictly shaped by the segmented or unsegmented nature of linguistic input. Instead, it suggests that children may bring a natural bias toward holistic learning, helping them to acquire language in versatile ways.
The researchers explored this process by focusing on homesigners—deaf children who, lacking exposure to spoken or sign language, create their own gestural communication systems. These homesign systems share many features of natural languages but are developed independently of a conventional language model. The study revealed that homesigners initially use gestures to represent entire meanings, with no breakdown into distinct components like specific hand shapes or motions. As they grow, however, these children begin to dissect these whole gestures, creating smaller, meaningful parts that they can recombine to convey nuanced meanings. Remarkably, this whole-to-part learning emerged even without linguistic input that typically encourages children to analyse language structure.
“Our findings indicate that children are not only responding to their environment but may come with an inherent preference for certain ways of learning language, including the ability to start with broad meanings and discover parts later,” Professor Arnon explains. “This has important implications for understanding the flexibility and resilience of language learning across different contexts. Children’s ability to grasp language by moving from wholes to parts may serve early communicative needs, allowing them to convey complex ideas before developing more intricate structures.”
The study’s findings suggest that whole-to-part learning is not just a fallback for situations where language input is difficult to segment; rather, it is an approach that children naturally bring to communication. This insight has broader implications, from understanding how languages evolve and adapt over generations to examining the developmental stages of emerging sign languages such as Nicaraguan Sign Language, which also exhibit a transition from holistic forms to more segmented, combinable parts.
Professor Arnon and Professor Goldin-Meadow’s research offers a fresh perspective on language acquisition, highlighting that children may begin with holistic expressions and gradually develop the ability to analyse and recombine parts. This approach underscores the adaptability of human language across diverse contexts, from spoken languages to homesign systems and beyond.
The research paper titled “Whole-to-part development in language creation” is now available in Trends in Cognitive Sciences and can be accessed HERE.