Despite notable exceptions, research on requests in world Englishes has so far largely involved role plays, questionnaires and discourse completion tasks. Moreover, research on requests in South Asian varieties of English is rather scarce. Therefore, the present study employs a multifactorial approach towards requests in Indian and Sri Lankan English as compared to their historical input variety British English by investigating the spoken components of the International Corpus of English, hence involving authentic, non-intuition-based empirical data. Based on a conditional inference tree and a random forest extended via the integration of interaction predictors, the present article concludes that quantitative differences in the realisation patterns of requests in British, Indian and Sri Lankan English can be observed.

中文翻译：

---

印度语和斯里兰卡英语的请求

尽管有明显的例外，但迄今为止，对世界英语请求的研究主要涉及角色扮演、问卷调查和话语完成任务。此外，对南亚英语变体的请求的研究相当缺乏。因此，本研究通过调查国际英语语料库的口语成分，采用多因素方法来处理印度和斯里兰卡英语的请求，并与历史输入的英国英语进行比较，从而涉及真实的、非直觉的经验数据。基于条件推理树和通过交互预测变量的整合扩展的随机森林，本文得出的结论是，可以观察到英国英语、印度英语和斯里兰卡英语的请求实现模式的数量差异。