Why Learners Privilege Word-Order Over Case-Marking: A Cross-Linguistic Meta-Analysis, New Data From Estonian, Finnish and Polish, And a Discriminative Learning Model. Psychological Review, 2019-2025

Ambridge, Ben (2025). Why Learners Privilege Word-Order Over Case-Marking: A Cross-Linguistic Meta-Analysis, New Data From Estonian, Finnish and Polish, And a Discriminative Learning Model. Psychological Review, 2019-2025. [Data Collection]. Colchester, Essex: UK Data Service. 10.5255/UKDA-SN-858052

LuCiD's mission is to transform our understanding of how children learn to talk, and deliver the scientific evidence needed to design effective interventions in early years education and healthcare.

Learning to use language to communicate effectively is hugely important for society. Many children enter school without the language skills they need to succeed in the classroom, and these early weaknesses in language and communication are a major predictor of educational and social inequality in later life. To tackle this problem, we need to know the answers to a number of questions: How do children learn language from what they see and hear? What do measures of children's brain activity tell us about what they know at different ages? How do differences between children and differences in their environments affect how children learn to talk? Answering these questions is a major challenge for researchers, but, in the first phase of LuCiD, we have made great strides towards meeting this challenge by bringing together researchers from a range of different research backgrounds and with a range of different research skills.

In its next phase, LuCiD will build on this success by coordinating three research streams in the UK and abroad.

STREAM 1: FROM VARIATION TO EXPLANATION: will take what we have discovered about word learning and grammatical development and use it to explain development in children with Developmental Language Disorder.

STREAM 2: FROM SIMPLE TO COMPLEX: will take what we have discovered about communicative development and use it to understand how different groups of children learn to use language to communicate in the more complicated real-world situations that they will encounter when they enter school.

STREAM 3: BEYOND 0-5: will build on LuCiD's 0-5 project - a study of 80 children's language learning across the first 5 years - by a) using the 0-5 data to understand how children's curiosity-based exploration shapes their word learning; b) using the 0-5 data to build individualized computer models of how particular children perform across different experiments and across development; and c) following the 0-5 children into school and determining how their preschool language abilities impact on the beginnings of their literacy development.

In this research, we will seek to understand language learning using a range of different methods. We will observe and record children in natural interaction as well as studying their language in more controlled experiments and using behavioural measures and correlations with brain activity (EEG). Combining information collected using these different methods will constrain the types of explanations that can be proposed; and using computer models to understand our results will help us to create more accurate and comprehensive theories of how children learn.

The next phase of LuCiD will also include a COMMUNICATIONS AGENDA, a TECHNOLOGY AGENDA and a CAPACITY BUILDING PROGRAMME.

In the COMMUNICATIONS AGENDA, we will work with our IMPACT CHAMPIONS to ensure that parents know how they can best help their children learn to talk, and to give healthcare and education professionals and policy-makers the information they need to create training and intervention programmes that are firmly rooted in the latest research findings.

In the TECHNOLOGY AGENDA, we will make the new tools and research designs that we have developed, and the new data that we have collected, available to other researchers and practitioners on an open access basis.

In the CAPACITY BUILDING PROGRAMME, we will train new researchers in the range of different methods used across the Centre, and in how to communicate their findings to parents, educational professionals and policy makers. This will ensure the long-term future of language development research in the UK and of our approach to understanding how children learn to talk.

Data description (abstract)

Language acquisition is one of the crowning achievements of our species; though a long-standing and unresolved question is why many learners struggle with a particular core and fundamental sentence type. In English, a two-participant sentence like The dog chased the cat can mean only one thing. But in many languages worldwide, the meaning of the sentence can be flipped (e.g., to “the cat chased the dog”) even while the word-order stays the same, by the use of case-marked nouns (e.g., cat + NOMINATIVE, dog + ACCUSATIVE).

First (Study 1), we confirm using meta-analysis that, across 21 languages, children show a significant disadvantage for noncanonical (“flipped”) versus canonical (“unflipped”) sentences (Cohen’s d = 0.69, SE = 0.09).

Second (Study 2), we demonstrate with a new comprehension study of 3- to 6-year-old native learners of Estonian (N = 87), Finnish (N = 83), and Polish (N = 82) that this effect is not due to (a) lack of knowledge of individual inflected words, (b) lack of linguistic context, (c) use of still pictures as opposed to animations, or (d) a focus on relatively young children.

Third (Study 3), we show that this effect falls naturally out of a simple discriminative-learning model, in which the cue of case-marking struggles to compete with the more frequently available cue of word-order to predict which noun (e.g., dog and cat) is the AGENT (e.g., the chaser), and which the PATIENT (e.g., the one chased). Together, these findings demonstrate that languages which, on the surface, appear to be very different are learned in a similar way: by discriminating cues to meaning.

Downloads

Altmetric