A Survey of Code-switching: Linguistic and Social Perspectives for Language Technologies

TL;DR

This survey reviews linguistic and social aspects of code-switching, highlighting gaps in language technology tools and datasets, and aims to foster collaboration between linguists and computational scientists.

Contribution

It bridges linguistic insights on code-switching with computational challenges, emphasizing the need for better data, evaluation benchmarks, and models that incorporate sociolinguistic factors.

Findings

Massive language models struggle with diverse C-S types due to data gaps.

Current benchmarks for C-S are insufficient for multilingual evaluation.

There is a lack of end-to-end systems capturing sociolinguistic aspects of C-S.

Abstract

The analysis of data in which multiple languages are represented has gained popularity among computational linguists in recent years. So far, much of this research focuses mainly on the improvement of computational methods and largely ignores linguistic and social aspects of C-S discussed across a wide range of languages within the long-established literature in linguistics. To fill this gap, we offer a survey of code-switching (C-S) covering the literature in linguistics with a reflection on the key issues in language technologies. From the linguistic perspective, we provide an overview of structural and functional patterns of C-S focusing on the literature from European and Indian contexts as highly multilingual areas. From the language technologies perspective, we discuss how massive language models fail to represent diverse C-S types due to lack of appropriate training data, lack of robust evaluation benchmarks for C-S (across multilingual situations and types of C-S) and lack of end-to-end systems that cover sociolinguistic aspects of C-S as well. Our survey will be a step towards an outcome of mutual benefit for computational scientists and linguists with a shared interest in multilingualism and C-S.