Scholarlink

<p>When reading aloud mixed-language paragraphs, bilinguals occasionally produce <i>language intrusions</i>, that is, they translate written switch words to avoid producing them overtly. Intrusions provide evidence about the nature of <i>language control</i>, which refers to how bilinguals maintain control over which language they speak. The present study compared eye movements for successful versus unsuccessful switches during reading aloud of mixed-language paragraphs to reveal the cognitive mechanisms underlying language control failures. Forty-eight English-dominant Spanish-English bilinguals read aloud four stories four times (once each in English-only, Spanish-only, English mixed-language, and Spanish mixed-language conditions). Mixed-language paragraphs included ten language switches (five content words, five function words). Though they elicited the most intrusion errors, dominant-language function (but not content) words elicited more frequent regressive eye movements on successful switches versus intrusions, indicating <i>heightened</i> monitoring of the most error-prone switch targets. Additionally, there was no evidence that inattention to dominant-language switches drove their relatively larger elicitation of intrusions – bilinguals skipped and gazed at switch words to similar extents in both languages, and regressed more often to dominant-language than nondominant-language switches regardless of whether they switched successfully or not. These results confirm the vulnerability of dominant-language function words to language control failures, and reveal that these failures occur despite <i>increased</i>, but ultimately only partially successful, attempted monitoring. More broadly, these results reinforce proposals that different retrieval mechanisms underlie production of function versus content words, but challenge claims that function words cannot be monitored at the level of planned production prior to production of overt speech.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Language switches can be monitored but not fully controlled: Eye-tracking evidence for syntax-driven language control

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月7日

查看全文
  • Jessie Quinn,
  • Victor S. Ferreira,
  • Tamar H. Gollan

摘要

When reading aloud mixed-language paragraphs, bilinguals occasionally produce language intrusions, that is, they translate written switch words to avoid producing them overtly. Intrusions provide evidence about the nature of language control, which refers to how bilinguals maintain control over which language they speak. The present study compared eye movements for successful versus unsuccessful switches during reading aloud of mixed-language paragraphs to reveal the cognitive mechanisms underlying language control failures. Forty-eight English-dominant Spanish-English bilinguals read aloud four stories four times (once each in English-only, Spanish-only, English mixed-language, and Spanish mixed-language conditions). Mixed-language paragraphs included ten language switches (five content words, five function words). Though they elicited the most intrusion errors, dominant-language function (but not content) words elicited more frequent regressive eye movements on successful switches versus intrusions, indicating heightened monitoring of the most error-prone switch targets. Additionally, there was no evidence that inattention to dominant-language switches drove their relatively larger elicitation of intrusions – bilinguals skipped and gazed at switch words to similar extents in both languages, and regressed more often to dominant-language than nondominant-language switches regardless of whether they switched successfully or not. These results confirm the vulnerability of dominant-language function words to language control failures, and reveal that these failures occur despite increased, but ultimately only partially successful, attempted monitoring. More broadly, these results reinforce proposals that different retrieval mechanisms underlie production of function versus content words, but challenge claims that function words cannot be monitored at the level of planned production prior to production of overt speech.