Abstract
It is common knowledge that grammar instruction is a prime factor in developing EFL students' proficiency. However, the argument revolves around the best method of teaching grammar, whether implicitly or explicitly. Metalanguage (ML), which stands for the technical terms that describe language, is essential for explicit instruction. As this teaching method is widely acknowledged, much research has focused on the relation between ML and language acquisition and proficiency. The focus has always been on the role of ML in developing students' language skills and uptake when taught through such a method. However, there is a relative paucity of studies that explore the exact impact of ML-based instruction on students' grammatical recognition, especially in the Arabian EFL context. Accordingly, the current study investigated the issue through an experimental method. The participants' (n = 73, 35 in the experimental group, and 38 in the control group) scores on the pre- and post-tests were analysed using t-test and item-analysis to test the research hypothesis. The results revealed that ML positively affects students' grammatical recognition. However, this improvement was evident only in some grammatical structures. These results imply that ML is beneficial for grammar instruction; however, a mixture of teaching methods should be applied to account for all grammatical constructs. The study findings may contribute to enriching the literature on ML and provide evidence for its importance in grammar instruction.