From Parallels to Divergence: Language Models' Learning of Action Nominal Constructions

查看原文 / View Original 编辑 / Edit

Computer Science > Computation and Language arXiv:2610.22468 (cs) [Submitted on 15 Oct 2026] Title: From Parallels to Divergence: Language Models’ Learning of Action Nominal Constructions Authors: Iskar Deng, Shane Steinert-Threlkeld.

计算机科学 > 计算与语言 arXiv:2610.22468 (cs) [提交于 2026 年 10 月 15 日] 标题：从平行到分歧：语言模型对动作名词结构的习得作者：Iskar Deng, Shane Steinert-Threlkeld。

Abstract: Action nominal constructions (ANCs) express events in nominal form, with their argument marking following patterns found in clauses or ordinary noun phrases. In this paper, we investigate how different grammatical configurations affect language models’ learning of ANCs and related constructions, and how these learning preferences compare with typological distributions in natural languages.

摘要：动作名词结构（ANCs）以名词形式表达事件，其论元标记遵循从句或普通名词短语中的模式。在本文中，我们研究了不同的语法配置如何影响语言模型对 ANC 及相关结构的习得，以及这些习得偏好与自然语言中类型学分布的对比情况。

We use a grammar-engineered method for synthetic corpus construction that combines shared semantic input from natural corpora with computational grammars. We construct 96 target languages varying in clause word order, noun-phrase word order, alignment, complement system, and ANC strategy.

我们使用一种语法工程方法构建合成语料库，将来自自然语料库的共享语义输入与计算语法相结合。我们构建了 96 种目标语言，这些语言在从句语序、名词短语语序、配列、补语系统和 ANC 策略上各不相同。

For each language, we train GPT-2-small models from scratch and evaluate them using targeted minimal pairs covering 34 grammatical phenomena. The results show typological parallels in basic clause configurations but divergences in complement systems and ANC strategies, with models also learning combinations rare or unattested in natural languages.

对于每种语言，我们从头开始训练 GPT-2-small 模型，并使用涵盖 34 种语法现象的针对性最小对进行评估。结果显示，基本从句配置存在类型学上的平行性，但在补语系统和 ANC 策略上存在分歧，模型还习得了自然语言中罕见或未证实的组合。

This partial alignment suggests that natural language distributions may reflect both general learning pressures and functional demands not fully captured by model learning experiments.

这种部分一致性表明，自然语言的分布可能既反映了普遍的学习压力，也反映了模型学习实验未能完全捕捉到的功能需求。