This study reports evidence of cross-linguistic influence (CLI) that surfaced from English compositions of SiSwati learners of English in Swaziland, where English is a second language. Although CLI has been studied widely in other languages, it has not been studied in SiSwati and English, and its implications for instruction are not known. Specifically, this study examined cases of negative transfer from the former to the latter, by identifying the cognitive influence participants' knowledge of their first language (L1) exerted on the structural acquisition of their second language (L2); how the knowledge and command of their L1 thwarts the process of learning L2, and the overall implications of this phenomenon on teaching practices. A total of sixty (60) narrative compositions from thirty (30) participants were collected, transcribed, and analyzed using contrastive rhetoric and categorical aggregation methods to establish consistency of the structural patterns in L2 learners' performance. In order to construct a psycholinguistic path that learners of English in Swaziland traverse during the acquisition of an L2, a weak version of contrastive analysis (CA) was used. The results revealed errors in the use of verbs and subject-verb agreement (SVO); however, a lot of errors were a function of lexical and structural transfer. Overall, this study is useful in improving language instruction in Swaziland and other similar ESL contexts.