<p>Monolingual bias remains a powerful force in Taiwan’s English language education, where learners are often encouraged to suppress their first language (L1) in pursuit of “authentic” English. This approach rests on outdated assumptions that view L1 knowledge as interference rather than a resource. Drawing on cognitive linguistics, this article advances a theory-driven reframing of the L1–L2 relationship, positioning Mandarin not as an obstacle but as a cognitive scaffold for English learning. The argument integrates three grammatical domains—(a) the perfect aspect, (b) comparative structures, and (c) the middle voice—under a unifying conceptual architecture that identifies shared cognitive mechanisms: <i>metaphorical possession grounding</i> (treating experiences as “possessed”), <i>metonymic tolerance</i> (flexibility in meaning extension), and <i>figure–ground alignment</i> (a construal strategy that backgrounds or omits the agent). Recognising these cognitive parallels allows educators to tap into learners’ existing conceptual resources, thereby challenging deficit-based views of non-native speakers and supporting diverse linguistic identities. While grounded in Taiwan’s English as a Foreign Language (EFL) context, the framework has wider relevance for multilingual education policy and for teacher training in contexts where the L1 and L2 are typologically distant. Keywords: cognitive linguistics; English as a Foreign Language (EFL); cross-linguistic influence; language teaching; monolingual bias; L1–L2 interface</p>