This article investigates how lexical cohesion is realised in Ewe vis-à-vis English. It also seeks to find out the constraints that inform the choice of lexical cohesion in Ewe-English bilingual constructions (code-switching). Employing both written and spoken data among undergraduate students who are native speakers of Ewe, the findings reveal that the Ewe language mirrors English perfectly in its realisation of lexical cohesion. It follows that Ewe realises lexical cohesion by repetition, simple and complex synonymy, hyponymy, juxtaposition and the like just as English does, except that Ewe is more phrasal in realising cohesion lexically. In addition, constraints such as paraphrasing and complementing were discovered as the constraints that affect the choice of lexical items in Ewe-English code-switching. What is noteworthy about these constraints is that they are not matrix language or embedded language dependent; the constraints are neutral and could affect either language of the Ewe-English bilingual in code-switching. This study therefore corroborates earlier studies that either language of the bilingual, matrix or embedded is capable of controlling the grammar of bilingual constructions.

本文研究了如何用Ewe相对于英语实现词汇衔接。它还试图找出在Ewe-English双语构造（代码转换）中选择词法衔接的约束条件。研究结果表明，利用母羊母语的本科生的书面和口头数据，母羊语言在词法衔接方面完美地反映了英语。随之而来的是，Ewe像英语一样通过重复，简单和复杂的同义词，下位，并列等来实现词汇衔接，只是Ewe在词法上实现衔接方面更具有短语意义。此外，还发现诸如释义和补语之类的约束是影响Ewe-英语代码转换中词汇项目选择的约束。这些约束值得注意的是它们与矩阵语言或嵌入式语言无关。这些限制是中性的，可能会影响代码转换中的Ewe-English双语。因此，这项研究证实了早期的研究，即双语，矩阵语言或嵌入式语言都能够控制双语构造的语法。