@article{kunz-2026-preferences,
title = "Preferences for Idiomatic Language are Acquired Slowly {---} and Forgotten Quickly: A Case Study on {S}wedish",
author = "Kunz, Jenny",
journal = "Transactions of the Association for Computational Linguistics",
volume = "14",
year = "2026",
address = "Cambridge, MA",
publisher = "MIT Press",
url = "https://aclanthology.org/2026.tacl-1.57/",
doi = "10.1162/tacl.a.727",
pages = "1266--1285",
abstract = "In this study, we investigate how language models develop preferences for idiomatic as compared to linguistically acceptable Swedish, both during pretraining and when adapting a model from English to Swedish. To do so, we train models on Swedish from scratch and by fine-tuning English-pretrained models, probing their preferences at various checkpoints using minimal pairs that differ in linguistic acceptability or idiomaticity. For linguistic acceptability, we adapt existing benchmarks into a minimal-pair format. To assess idiomaticity, we introduce two novel datasets: one contrasting conventionalized idioms with plausible variants, and another contrasting idiomatic Swedish with Translationese. Our findings suggest that idiomatic competence emerges more slowly than other linguistic abilities, including grammatical and lexical correctness. While longer training yields diminishing returns for most tasks, idiom-related performance continues to improve, particularly in the largest model tested (8B). However, instruction tuning on data machine-translated from English {---} the common approach for languages with little or no native instruction data {---} causes models to rapidly lose their preference for idiomatic language."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kunz-2026-preferences">
<titleInfo>
<title>Preferences for Idiomatic Language are Acquired Slowly — and Forgotten Quickly: A Case Study on Swedish</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jenny</namePart>
<namePart type="family">Kunz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<genre authority="bibutilsgt">journal article</genre>
<relatedItem type="host">
<titleInfo>
<title>Transactions of the Association for Computational Linguistics</title>
</titleInfo>
<originInfo>
<issuance>continuing</issuance>
<publisher>MIT Press</publisher>
<place>
<placeTerm type="text">Cambridge, MA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">periodical</genre>
<genre authority="bibutilsgt">academic journal</genre>
</relatedItem>
<abstract>In this study, we investigate how language models develop preferences for idiomatic as compared to linguistically acceptable Swedish, both during pretraining and when adapting a model from English to Swedish. To do so, we train models on Swedish from scratch and by fine-tuning English-pretrained models, probing their preferences at various checkpoints using minimal pairs that differ in linguistic acceptability or idiomaticity. For linguistic acceptability, we adapt existing benchmarks into a minimal-pair format. To assess idiomaticity, we introduce two novel datasets: one contrasting conventionalized idioms with plausible variants, and another contrasting idiomatic Swedish with Translationese. Our findings suggest that idiomatic competence emerges more slowly than other linguistic abilities, including grammatical and lexical correctness. While longer training yields diminishing returns for most tasks, idiom-related performance continues to improve, particularly in the largest model tested (8B). However, instruction tuning on data machine-translated from English — the common approach for languages with little or no native instruction data — causes models to rapidly lose their preference for idiomatic language.</abstract>
<identifier type="citekey">kunz-2026-preferences</identifier>
<identifier type="doi">10.1162/tacl.a.727</identifier>
<location>
<url>https://aclanthology.org/2026.tacl-1.57/</url>
</location>
<part>
<date>2026</date>
<detail type="volume"><number>14</number></detail>
<extent unit="page">
<start>1266</start>
<end>1285</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Journal Article
%T Preferences for Idiomatic Language are Acquired Slowly — and Forgotten Quickly: A Case Study on Swedish
%A Kunz, Jenny
%J Transactions of the Association for Computational Linguistics
%D 2026
%V 14
%I MIT Press
%C Cambridge, MA
%F kunz-2026-preferences
%X In this study, we investigate how language models develop preferences for idiomatic as compared to linguistically acceptable Swedish, both during pretraining and when adapting a model from English to Swedish. To do so, we train models on Swedish from scratch and by fine-tuning English-pretrained models, probing their preferences at various checkpoints using minimal pairs that differ in linguistic acceptability or idiomaticity. For linguistic acceptability, we adapt existing benchmarks into a minimal-pair format. To assess idiomaticity, we introduce two novel datasets: one contrasting conventionalized idioms with plausible variants, and another contrasting idiomatic Swedish with Translationese. Our findings suggest that idiomatic competence emerges more slowly than other linguistic abilities, including grammatical and lexical correctness. While longer training yields diminishing returns for most tasks, idiom-related performance continues to improve, particularly in the largest model tested (8B). However, instruction tuning on data machine-translated from English — the common approach for languages with little or no native instruction data — causes models to rapidly lose their preference for idiomatic language.
%R 10.1162/tacl.a.727
%U https://aclanthology.org/2026.tacl-1.57/
%U https://doi.org/10.1162/tacl.a.727
%P 1266-1285
Markdown (Informal)
[Preferences for Idiomatic Language are Acquired Slowly — and Forgotten Quickly: A Case Study on Swedish](https://aclanthology.org/2026.tacl-1.57/) (Kunz, TACL 2026)
ACL