This dissertation examines metaphorical language use in Georgian EFL learners’ argumentative essays and the role of proficiency level in metaphor production, given the absence of prior research in the Georgian EFL context. Case Study 1 evaluates existing metaphor identification methods and justifies the selection of the Metaphor Identification Procedure Vrije Universiteit (MIPVU) for use in this study, demonstrating that it captures a wider range of both conventional and creative metaphorical expressions in L2 writing. Case Study 2 analyzes EFL argumentative essays written by Georgian learners to (i) measure metaphor density across various proficiency levels, (ii) identify categories of metaphorical errors, and (iii) examine crosslinguistic influences on these errors. The analysis reveals increased metaphor production at higher proficiency levels, with prepositions emerging as the most frequently used metaphor-related words. Error categorization identifies level-specific challenges, with semantic and grammatical difficulties characterizing the intermediate level and a higher frequency of phrasal errors at the advanced level. L1 influence is most evident at the B2 level, where learners show a greater tendency to experiment with metaphorical language. Case Study 3 explores the effects of teaching metaphorical expressions in the L2 classroom by applying Conceptual Metaphor Theory (CMT) in a four-week study with advanced-level learners, representing the first empirical investigation of this approach with Georgian EFL learners. Analysis of data from the experiment reveals significant short- and long-term improvements in learners’ ability to produce metaphors following explicit instruction on conceptual metaphors, and further indicates that integrating CMT with the task-based language teaching method (TBLT) supports sustained metaphor retention and production at advanced proficiency levels. Together, the three case studies offer valuable theoretical, practical, and pedagogical implications for metaphorical language production in L2. These implications extend beyond the context of Georgian EFL learners, as they demonstrate how learner proficiency and L1 influence affect metaphor use in L2 writing and provide insights that can inform research and teaching more broadly.