Abstract Adopting a usage-based perspective, the present study assesses the mental lexicon of Turkish heritage speakers in the UK (HSs, n = 31) regarding the productive use of formulaic inflectional suffix templates and the level of sophistication of the lemmas produced in free speech. We additionally explore input-related predictors of this performance by comparing HS performance to that of a group of previous generation immigrant bilinguals (IBs, n = 61), who are representative input providers, and of a group of monolinguals (n = 44). The results show that overall, both the HSs and IBs diverge from the monolinguals in that they use nominal suffix sequences less productively and rely on less sophisticated nominal lemmas. Their verbal productivity performance, however, remains intact. We argue that altered input results in a performance which diverges from that of the monolinguals but converges on the immigrant variety. The individual variability is partly explained by the amount of L1 passive exposure, indicating that the HSs are not sensitive only to the changes in the input available to them but also to the amount of it. These findings provide new insights into the line of research that describes HSs as incomplete due to their L1-divergent skills in comparison to a monolingual baseline.

中文翻译：

---

一种基于用法的方法，用于在成人遗产演讲者的表演中有效地使用屈折模式和引理复杂程度

摘要 本研究采用基于使用的观点，评估了英国土耳其传统演讲者（HS，n = 31）的心理词汇，关于公式化屈折后缀模板的生产性使用和言论自由中产生的引理的复杂程度. 我们还通过将 HS 表现与作为代表性输入提供者的一组上一代移民双语者（IB，n = 61）和一组单语者（n = 44）的表现进行比较，探索了与输入相关的这种表现的预测因素。结果表明，总体而言，HS 和 IB 都与单语者不同，因为它们使用标称后缀序列的效率较低，并且依赖不太复杂的标称引理。然而，他们的口头生产力表现保持不变。我们认为，改变的输入会导致表现与单一语言者的表现不同，但会收敛于移民的表现。个体差异部分是由 L1 被动暴露的数量解释的，这表明 HS 不仅对他们可用的输入变化敏感，而且对其数量也很敏感。这些发现为将 HS 描述为不完整的研究线提供了新的见解，因为与单语基线相比，它们的 L1 发散技能。