The adaptive control hypothesis developed by Green and Abutalebi is the most influential theory of bilingual language control. The focus of this article is on the predictions that other researchers have derived based on the three different modes of interactional context described by the hypothesis. Foremost, that dual-language contexts should enhance domain-general executive functions more than single-language contexts. Several recent and ambitious behavioral tests of these predictions are reviewed. Although there was some evidence that dual-language contexts are associated with smaller switch costs, the evidence is inconsistent and there were no similar advantages for inhibitory control. The hypothesis also predicts neuroanatomical adaptations to the three types of interactional context. A careful evaluation of the relevant fMRI and ERP studies that take into account whether behavioral differences align with neuroscience differences and resolves valence ambiguities led to the conclusion that the neuroscience evidence for the hypothesis is, at best, inconsistent. The study also includes new analyses of two large-sample studies that enable the identification of relatively pure cases of single-language bilinguals, dual-language bilinguals, and dense-code switchers. Across nine different measures of executive functioning, the predicted advantage of the dual-language context never materialized. The hypotheses derived from the adaptive control hypothesis do not accurately predict behavioral performance on tests of executive functioning and do not advance our understanding as to what dimensions of bilingualism may lead to enhancements in specific components of executive functioning.

中文翻译：

---

由自适应控制假说启发的关于语言转换对执行功能的认知后果的预测经常失败

Green 和 Abutalebi 提出的自适应控制假说是双语语言控制最有影响的理论。本文的重点是其他研究人员根据假设描述的三种不同交互上下文模式得出的预测。最重要的是，双语语境比单语语境应该更能增强域一般执行功能。回顾了最近对这些预测进行的一些雄心勃勃的行为测试。虽然有一些证据表明双语环境与较小的转换成本相关，但证据不一致，抑制控制没有类似的优势。该假设还预测了对三种交互环境的神经解剖学适应。对相关 fMRI 和 ERP 研究进行仔细评估，这些研究考虑了行为差异是否与神经科学差异一致并解决了价数歧义，得出的结论是，该假设的神经科学证据充其量是不一致的。该研究还包括对两项大样本研究的新分析，这些研究能够识别相对纯的单语双语者、双语双语者和密集代码切换者的案例。在九种不同的执行功能测量中，双语环境的预测优势从未实现。