@inproceedings{hwang-2026-child,
title = "Can Child Language Acquisition be Modeled by Linear Causal or Bidirectional Masked Models Trained on Child-Directed Input?: A Comparison of Baby Language Models and Children in Acquiring Wanna Contraction",
author = "Hwang, Haerim",
editor = "Choshen, Leshem and
Cotterell, Ryan and
Gul, Mustafa Omer and
Jumelet, Jaap and
Linzen, Tal and
Mueller, Aaron and
Salhan, Suchir and
Shah, Raj Sanjay and
Warstadt, Alex and
Wilcox, Ethan Gotlieb",
booktitle = "Proceedings of the Second {B}aby{LM} Workshop",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.babylm-main.5/",
pages = "50--64",
ISBN = "979-8-89176-475-0",
abstract = "Children{'}s remarkable ability to acquire language remains one of the most fascinating questions in cognitive science. How do they develop sophisticated syntactic knowledge? To contribute to this question, the present study investigates the role of input and learning mechanisms by modeling child language acquisition across increasing age. Specifically, this study trains 30 baby language models (LMs) on naturalistic input directed to children up to ages 3 through 7 and compares their syntactic knowledge directly with that of children aged 3 to 7, focusing on the constraint on wanna contraction, a phenomenon not previously tested in language modeling. Furthermore, the study examines two major LM architectures: an autoregressive causal LM, represented by GPT, and a bidirectional masked LM, represented by RoBERTa. The results showed that children were sensitive to the constraint and displayed age-related development, but neither GPT-based nor RoBERTa-based baby models showed the same pattern. In this respect, the baby LMs do not seem to fully capture human-like development, possibly suggesting that the learning mechanisms underlying child language acquisition do not entirely align with the architectures tested here and cannot be reduced to input-driven learning alone within the context of these models."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hwang-2026-child">
<titleInfo>
<title>Can Child Language Acquisition be Modeled by Linear Causal or Bidirectional Masked Models Trained on Child-Directed Input?: A Comparison of Baby Language Models and Children in Acquiring Wanna Contraction</title>
</titleInfo>
<name type="personal">
<namePart type="given">Haerim</namePart>
<namePart type="family">Hwang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Second BabyLM Workshop</title>
</titleInfo>
<name type="personal">
<namePart type="given">Leshem</namePart>
<namePart type="family">Choshen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ryan</namePart>
<namePart type="family">Cotterell</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mustafa</namePart>
<namePart type="given">Omer</namePart>
<namePart type="family">Gul</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaap</namePart>
<namePart type="family">Jumelet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tal</namePart>
<namePart type="family">Linzen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aaron</namePart>
<namePart type="family">Mueller</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Suchir</namePart>
<namePart type="family">Salhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Raj</namePart>
<namePart type="given">Sanjay</namePart>
<namePart type="family">Shah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alex</namePart>
<namePart type="family">Warstadt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ethan</namePart>
<namePart type="given">Gotlieb</namePart>
<namePart type="family">Wilcox</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-475-0</identifier>
</relatedItem>
<abstract>Children’s remarkable ability to acquire language remains one of the most fascinating questions in cognitive science. How do they develop sophisticated syntactic knowledge? To contribute to this question, the present study investigates the role of input and learning mechanisms by modeling child language acquisition across increasing age. Specifically, this study trains 30 baby language models (LMs) on naturalistic input directed to children up to ages 3 through 7 and compares their syntactic knowledge directly with that of children aged 3 to 7, focusing on the constraint on wanna contraction, a phenomenon not previously tested in language modeling. Furthermore, the study examines two major LM architectures: an autoregressive causal LM, represented by GPT, and a bidirectional masked LM, represented by RoBERTa. The results showed that children were sensitive to the constraint and displayed age-related development, but neither GPT-based nor RoBERTa-based baby models showed the same pattern. In this respect, the baby LMs do not seem to fully capture human-like development, possibly suggesting that the learning mechanisms underlying child language acquisition do not entirely align with the architectures tested here and cannot be reduced to input-driven learning alone within the context of these models.</abstract>
<identifier type="citekey">hwang-2026-child</identifier>
<location>
<url>https://aclanthology.org/2026.babylm-main.5/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>50</start>
<end>64</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Can Child Language Acquisition be Modeled by Linear Causal or Bidirectional Masked Models Trained on Child-Directed Input?: A Comparison of Baby Language Models and Children in Acquiring Wanna Contraction
%A Hwang, Haerim
%Y Choshen, Leshem
%Y Cotterell, Ryan
%Y Gul, Mustafa Omer
%Y Jumelet, Jaap
%Y Linzen, Tal
%Y Mueller, Aaron
%Y Salhan, Suchir
%Y Shah, Raj Sanjay
%Y Warstadt, Alex
%Y Wilcox, Ethan Gotlieb
%S Proceedings of the Second BabyLM Workshop
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-475-0
%F hwang-2026-child
%X Children’s remarkable ability to acquire language remains one of the most fascinating questions in cognitive science. How do they develop sophisticated syntactic knowledge? To contribute to this question, the present study investigates the role of input and learning mechanisms by modeling child language acquisition across increasing age. Specifically, this study trains 30 baby language models (LMs) on naturalistic input directed to children up to ages 3 through 7 and compares their syntactic knowledge directly with that of children aged 3 to 7, focusing on the constraint on wanna contraction, a phenomenon not previously tested in language modeling. Furthermore, the study examines two major LM architectures: an autoregressive causal LM, represented by GPT, and a bidirectional masked LM, represented by RoBERTa. The results showed that children were sensitive to the constraint and displayed age-related development, but neither GPT-based nor RoBERTa-based baby models showed the same pattern. In this respect, the baby LMs do not seem to fully capture human-like development, possibly suggesting that the learning mechanisms underlying child language acquisition do not entirely align with the architectures tested here and cannot be reduced to input-driven learning alone within the context of these models.
%U https://aclanthology.org/2026.babylm-main.5/
%P 50-64
Markdown (Informal)
[Can Child Language Acquisition be Modeled by Linear Causal or Bidirectional Masked Models Trained on Child-Directed Input?: A Comparison of Baby Language Models and Children in Acquiring Wanna Contraction](https://aclanthology.org/2026.babylm-main.5/) (Hwang, BabyLM 2026)
ACL